Return-Path: william@bourbon.usc.edu Delivery-Date: Fri May 11 22:18:42 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.2 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, HOT_NASTY,NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l4C5Ig2I008542 for ; Fri, 11 May 2007 22:18:42 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l4C5HHfS013237 for ; Fri, 11 May 2007 22:17:17 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l4C5HHBW013236 for cs551@merlot; Fri, 11 May 2007 22:17:17 -0700 Date: Fri, 11 May 2007 22:17:17 -0700 From: william@bourbon.usc.edu Message-Id: <200705120517.l4C5HHBW013236@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: updated final project part (2) statistics... Hi, After the regrade, the updated final project part (2) statistics is as follows: Count = 33 Avg = 46.41 StdDev = 26.89 Max = 84.00 Min = 1.00 1 80+ X 8 75+ XXXXXXXX 3 70+ XXX 0 65+ 2 60+ XX 0 55+ 0 50+ 0 45+ 3 40+ XXX 5 35+ XXXXX 0 30+ 1 25+ X 5 20+ XXXXX 2 15+ XX 0 10+ 0 5+ 3 0+ XXX The statics before the regrade was: Count = 33 Avg = 38.61 StdDev = 27.19 Max = 79.50 Min = 1.00 6 75+ XXXXXX 1 70+ X 0 65+ 2 60+ XX 0 55+ 4 50+ XXXX 0 45+ 2 40+ XX 2 35+ XX 2 30+ XX 2 25+ XX 5 20+ XXXXX 0 15+ 0 10+ 0 5+ 7 0+ XXXXXXX -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue May 8 23:18:08 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l496I8gV030306 for ; Tue, 8 May 2007 23:18:08 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l496GsBr031249 for ; Tue, 8 May 2007 23:16:54 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l496GsMX031248 for cs551@merlot; Tue, 8 May 2007 23:16:54 -0700 Date: Tue, 8 May 2007 23:16:54 -0700 From: william@bourbon.usc.edu Message-Id: <200705090616.l496GsMX031248@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Regrade for final project part (2) is this Thursday in SAL 211... Hi, I will e-mail your final project part (2) grade to you soon. Just want to mention that the grader, Pritam Pradhan , has reserved the whole Thursday (8:30am - 6pm) for regrade *appointments* in SAL 211. (There will be on "walk-ins".) I must turn in grades this Friday at the latest. So, please try your best to get the regrades started this Thursday. If you cannot, don't worry about it. I will still send in the grades this Friday. If there is a grade change, it can be done quite easily. The regrade procedure will be the same as regrades for final project part (1). Please make an appointment either on the hour or 30 minutes past the hour and please mention which of the following 3 categories you fall into: 1) If your score is 1 or very low, you should only make a 15-minute appointment to "get around the grading obstacle". 2) If not (1) and only asking to regrade one sub-section (as in (B)(1)), also make a 15-minute appointment. 3) If neither (1) nor (2), make a 30-minute appointment. Please give Pritam at least 2 timeslots so he can choose one. Please note that there is a *new rule* for the regrade (clearly, this is for minor change only since part (2) is worth 85 points total): You can modify your code at 25 points per line of change (as defined in the projects class web page) during regrade! Finally, just want to remind everyone that the final exam is tomorrow (Wednesday, 5/9) from 2pm to 4pm in THH 212. Please check the News (red) section of the web page for information about the final exam. See you tomorrow! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon May 7 21:02:27 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l4842Rce028720 for ; Mon, 7 May 2007 21:02:27 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l4841HwU009608 for ; Mon, 7 May 2007 21:01:17 -0700 Message-Id: <200705080401.l4841HwU009608@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About DSR Date: Mon, 07 May 2007 21:01:17 -0700 From: william@bourbon.usc.edu Someone wrote: > The route request routing in wireless network seems same as what > we did in final project. but I have a doubt a host discards the > request packet when it sees him self in the route recorded in the > packet. I don't see this condition mentioned in the slides. Could you tell me where this condition came from? > but when this condition will happen? because before > seeing the recorded route the host'll discard the packet as soon > as he sees it because he would have seen the same packet before > isn't it right? or I am missing something? -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon May 7 20:24:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l483OILA025760 for ; Mon, 7 May 2007 20:24:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l483N7AF009032 for ; Mon, 7 May 2007 20:23:07 -0700 Message-Id: <200705080323.l483N7AF009032@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 28 slide 33 Date: Mon, 07 May 2007 20:23:07 -0700 From: william@bourbon.usc.edu Someone wrote: > In slide 33 here the data get fdropped becoz of RPF or becoz of > Truncated RPF??? > > I believe TRPF. Wht do you think? I'm not sure what truncated RPF is. But I think RPF works here. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon May 7 13:45:15 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l47KjFoF026970 for ; Mon, 7 May 2007 13:45:15 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l47Ki61G004767 for ; Mon, 7 May 2007 13:44:06 -0700 Message-Id: <200705072044.l47Ki61G004767@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Inter-AS multicast Date: Mon, 07 May 2007 13:44:06 -0700 From: william@bourbon.usc.edu Someone wrote: > In face of the resistance of multicast from ISP, what is > the most commonly used inter-AS multicast scheme > at this moment? Is it DVMRP or do we have to resort to MBone? I don't know what the answer is. I don't keep up with multicast since I do not do research in this area. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun May 6 19:43:06 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l472h6Y3004100 for ; Sun, 6 May 2007 19:43:06 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l472fxZg025719 for ; Sun, 6 May 2007 19:41:59 -0700 Message-Id: <200705070241.l472fxZg025719@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About Bistro Date: Sun, 06 May 2007 19:41:59 -0700 From: william@bourbon.usc.edu Someone wrote: > But > If we cannot form a chain of trust from a client machine to a > dest Bistro then how can a user be satisfied that his data has > not been modified at the intermediate bistro. Coz then the dest > Bistro will reject the submission since the SHA value will > not match and the innocent user would face problems. What problem? The user has a digitally signed ticket which proved that a submission with a certain timestamp was submitted ontime. If the user can produce a document with the hash in the signed ticket, the Event Owner must accept this document. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Sunday, May 6, 2007 4:48 pm Subject: Re: About Bistro To: cs551@merlot.usc.edu > Someone wrote: > > > Also in slide 28. End operator trusts Dest Bistro and End User > > Trusts its client software and Dest Bistro. DOes it mean that > > the client software is the entity that trusts the Intermediate > > bistros?? Or does the client himself ??? > > No one trusts any of the intermediate bistros! > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sun May 6 16:48:28 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l46NmSEX022935 for ; Sun, 6 May 2007 16:48:28 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l46NlM0r024283 for ; Sun, 6 May 2007 16:47:22 -0700 Message-Id: <200705062347.l46NlM0r024283@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About Bistro Date: Sun, 06 May 2007 16:47:22 -0700 From: william@bourbon.usc.edu Someone wrote: > Also in slide 28. End operator trusts Dest Bistro and End User > Trusts its client software and Dest Bistro. DOes it mean that > the client software is the entity that trusts the Intermediate > bistros?? Or does the client himself ??? No one trusts any of the intermediate bistros! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun May 6 16:16:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l46NGINO020436 for ; Sun, 6 May 2007 16:16:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l46NFCUY023933 for ; Sun, 6 May 2007 16:15:12 -0700 Message-Id: <200705062315.l46NFCUY023933@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About Bistro Date: Sun, 06 May 2007 16:15:12 -0700 From: william@bourbon.usc.edu Someone wrote: > With reference to the earlier mail, the data transfer is, in a > sense, independent of the deadline.Does it mean that there > isnt any constraint on when the actual data is uploaded on any > intermediate bistro??? If the 2nd phase is not finished, the user may not be happy. In the worst case, the 2nd phase (and 3rd phase) is skipped. At the end, the event owner can send an e-mail to the user and ask for the data with the right hash. > Also in the slide 27 of lecture 26 after the first step is over > how does this client gets the SESSION KEY?? I believe that the > destination bistro sends the SESSION KEY in the > way....Kpriv[h(T),Email,s,sigma,Kses]???Right?? No. A random session key is generated by the client. The destination bistro does not have to know the session key until the end of the 3rd phase. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Sunday, May 6, 2007 3:52 pm Subject: Re: About Bistro To: cs551@merlot.usc.edu > Someone wrote: > > > Is the "commit" phase (client-push) transparent to client? > > I'm not sure what you meant by "transparent". The client > has to push the data to a bistro server (any bistro server) > on the Internet. > > > What happens if a Bistro server is overloaded with "push". > > I guess there will be a delay in upload of actual data (could > > be very large for each client). If yes, how to determine > > the length of the delay? > > Once the first phase is completed, the time it takes for the > 2nd (commit) phase does not matter as far as the correctness > of the protocol is concerned. If the 2nd phase takes too > long, the user might be upset that it's taking too long. The > user can turn off his/her machine and restart the 2nd phase > at any time (even after the deadline). So, the data transfer > is, in a sense, independent of the deadline. > > > Also, is it required that Bistro-like system be installed > > on intermediary systems. Can they be installed on > > destination server if scalability is OK? > > Sure! For our homework/project submission, the data is > uploaded to the destination bistro directly during the 2nd > phase. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sun May 6 15:52:10 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l46MqA1o018564 for ; Sun, 6 May 2007 15:52:10 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l46Mp4oi023549 for ; Sun, 6 May 2007 15:51:04 -0700 Message-Id: <200705062251.l46Mp4oi023549@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About Bistro Date: Sun, 06 May 2007 15:51:04 -0700 From: william@bourbon.usc.edu Someone wrote: > Is the "commit" phase (client-push) transparent to client? I'm not sure what you meant by "transparent". The client has to push the data to a bistro server (any bistro server) on the Internet. > What happens if a Bistro server is overloaded with "push". > I guess there will be a delay in upload of actual data (could > be very large for each client). If yes, how to determine > the length of the delay? Once the first phase is completed, the time it takes for the 2nd (commit) phase does not matter as far as the correctness of the protocol is concerned. If the 2nd phase takes too long, the user might be upset that it's taking too long. The user can turn off his/her machine and restart the 2nd phase at any time (even after the deadline). So, the data transfer is, in a sense, independent of the deadline. > Also, is it required that Bistro-like system be installed > on intermediary systems. Can they be installed on > destination server if scalability is OK? Sure! For our homework/project submission, the data is uploaded to the destination bistro directly during the 2nd phase. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri May 4 22:32:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l455WXsI024104 for ; Fri, 4 May 2007 22:32:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l455VXPa008147 for ; Fri, 4 May 2007 22:31:33 -0700 Message-Id: <200705050531.l455VXPa008147@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Lecture 24 Date: Fri, 04 May 2007 22:31:33 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a doubt regarding your previous reply, > > I believe Collision is very much different from the case of > "Hidden terminal" Coz in collison B cannot hear from either of > the two sources but in "hidden terminal" A is already sending > some byte stream to B when C tries to send to B. > > This is further exemplified in paper [MACAW] section 2.2 second > paragraph. > > Please correct If I am wrong Section 2.2 of the MACAW paper did say that "this produces a collision at B" (in the last paragraph on the 2nd page of the paper). The reason for the collision is "hidden terminal". In your original e-mail, you used the word "intervene" and it's not a technically word in this context. For the exams, if you use a word that does not have the right meaning, you won't get credit for it. You cannot argue that you mean the right thing but just put down the wrong word. To be fair to all, I can only grade what you wrote and not what's in your mind! -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Friday, May 4, 2007 8:21 pm Subject: Re: Lecture 24 To: cs551@merlot.usc.edu > Someone wrote: > > > 1) In lecture 24 slide 20 The "hidden terminal" Problem means A > > sending to B while C interveans. > > Correct?? > > Well, you should get the terminology correct. If you go to the > previous page, this is "collision". > > > 2) Also just wanted to confirm that . All the CTS packets are > > sent to all members in the Cell. > > Correct?? > > I wouldn't say "send to". "Send to" sounds like the case > that if you have N other nodes, you send N packets. > > These nodes are using wireless radios. One node puts out a > signal. Any node that's in range can hear it. > > > 3)Can it be interpretted that the CTS packets are multicasted > > ones?? > > Multicast is basically a mode of "addressing". It's more like > broadcast for wireless. Only nodes within range can hear the > broadcast. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Fri May 4 22:17:14 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l455HENn022857 for ; Fri, 4 May 2007 22:17:14 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l455GFAd007957 for ; Fri, 4 May 2007 22:16:15 -0700 Message-Id: <200705050516.l455GFAd007957@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Lecture 24 Date: Fri, 04 May 2007 22:16:15 -0700 From: william@bourbon.usc.edu Someone wrote: > Now In lecture 24 slide 24 when we add the ACK in link layer. > I am not seeing any " DS" here before A sends Data. > Am I interpretting it wrongly??? DS (from slide 22) is there. It's just not repeated here. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri May 4 20:20:54 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l453Kswv013675 for ; Fri, 4 May 2007 20:20:54 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l453Jtvx006637 for ; Fri, 4 May 2007 20:19:55 -0700 Message-Id: <200705050319.l453Jtvx006637@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Lecture 24 Date: Fri, 04 May 2007 20:19:55 -0700 From: william@bourbon.usc.edu Someone wrote: > 1) In lecture 24 slide 20 The "hidden terminal" Problem means A > sending to B while C interveans. > Correct?? Well, you should get the terminology correct. If you go to the previous page, this is "collision". > 2) Also just wanted to confirm that . All the CTS packets are > sent to all members in the Cell. > Correct?? I wouldn't say "send to". "Send to" sounds like the case that if you have N other nodes, you send N packets. These nodes are using wireless radios. One node puts out a signal. Any node that's in range can hear it. > 3)Can it be interpretted that the CTS packets are multicasted > ones?? Multicast is basically a mode of "addressing". It's more like broadcast for wireless. Only nodes within range can hear the broadcast. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu May 3 22:46:52 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l445kqKv008791 for ; Thu, 3 May 2007 22:46:52 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l445jtRa025596 for ; Thu, 3 May 2007 22:45:55 -0700 Message-Id: <200705040545.l445jtRa025596@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Snoop operation Date: Thu, 03 May 2007 22:45:55 -0700 From: william@bourbon.usc.edu Someone wrote: > What I mean is that whenever a SNOOP device receives > a new packet out of order, how can it immediately determine > that it's caused by congestion and not by the wireless > environment? Slide 34 of lecture 24 is about "FH-to-MH Snoop Data Processing". It's assume that there is no wireless connection from FH all the way to the basestation (running Snoop). The only wireless connection is from the basestation to the MH. So, if a data packet from FH to MH skip a sequence number, it's assume that the packet before was lost in the Internet (and all packet loss in the Internet are assumed to due to congestion). -- Bill Cheng // bill.cheng@usc.edu On 5/3/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > On p. 34 of Lec. 24, what does "greater than last ACKed, > > pass on" correspond to in the flowchart on the left? > > This corresponds to the "No" arrow out of "In-sequence". > This is the case of a congestion loss. > > > Also, in the flowchart, if a new packet is out of sequence, > > how could we immediately say that it is "congestion loss" > > (the red font under the box)? I think the idea of snoop > > is to abate TCP's overreaction to packet loss (some > > caused by wireless environment not congestion). > > "Congestion loss" means that a packet is dropped by a > congested router in the Internet. This is what TCP assumes > when any packet is lost. > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu May 3 22:39:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l445dX6S008149 for ; Thu, 3 May 2007 22:39:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l445cbAT025486 for ; Thu, 3 May 2007 22:38:37 -0700 Message-Id: <200705040538.l445cbAT025486@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: about two-bit diff-serv Date: Thu, 03 May 2007 22:38:37 -0700 From: william@bourbon.usc.edu Someone wrote: > 1.the two-bit diff-serv means Premium and Assured service, > setting the P bit and A bit? Correct. > 2.and in silde 13 lecture 22, "wait for token"&"Test if token" is > to check the profile is P or A? The top picture is for Premium traffic and the bottom picture is for Assured traffic. > but what's the difference for "wait for token"&"Test if token"? In the case of "test if token", the a packet is never delayed. When a packet arrives, if there is a token, the A bit is set; if there isn't, it's unmarked. In the case of "wait for token", if there is no token, the packet is held until a token arrives. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu May 3 22:35:20 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l445ZKCn007830 for ; Thu, 3 May 2007 22:35:20 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l445YNPU025436 for ; Thu, 3 May 2007 22:34:23 -0700 Message-Id: <200705040534.l445YNPU025436@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 21 slide no 27 Date: Thu, 03 May 2007 22:34:23 -0700 From: william@bourbon.usc.edu Someone wrote: > In the graph shown in slide everything seems to be perfectly fine > for flow A. How ever as the token bucket takes 1 sec to fill up > for the flow B we see a lower bandwidth than 1 innitially. > > Now I have a doubt here . Shoudnt the step up in the graph start > from 1sec instead of starting from 2 sec for flow B??Is it since > u are imagining your simulation to start from 1sec?? > > Also as the tocken bucket doesnt allow packet to flow until > enough packets accumulate Shoudnt the flow lines A and B start > from a 0 value and immediately jump to the desired value??? You should think of the drawing as a snapshot of something that has been running for a while. There is no simulation going on here. It's just a drawing to illustrate an idea. (I drew the picture myself and didn't try to match anything.) The idea here is that flow B can sustain a burst while flow A cannot. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu May 3 22:28:36 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l445Sath007341 for ; Thu, 3 May 2007 22:28:36 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l445ReGY025365 for ; Thu, 3 May 2007 22:27:40 -0700 Message-Id: <200705040527.l445ReGY025365@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 16 Date: Thu, 03 May 2007 22:27:40 -0700 From: william@bourbon.usc.edu Someone wrote: > I am just wondering, do we have to remember all the equations > provided in a lecture ? Are you going to ask to write an algorithm given > in slide 37 ? This slide is the RED pseudocode. Some of the equations are explained in the slides 38 and 40. I will *not* ask you to write any pseudocode for the final exam. I don't think you need to *memorize* any equation. If you understand what's going on with an equation, you should be able to derive the important part of the equation. I'm not saying that you don't have to memorize anything. For example, in physics, the force of gravity is inversely proportional to the square of distance. So, you have to memorize the "square" part. So, you have to memorize something, and that's okay because that's part of the "understanding". (I hope this is a good analogy of something.) Also, not all equations are equally important. I tend to ask questions about what I think are important equations, if I ask questions about equations at all. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu May 3 22:14:02 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l445E2kn006242 for ; Thu, 3 May 2007 22:14:02 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l445D6Xh025247 for ; Thu, 3 May 2007 22:13:06 -0700 Message-Id: <200705040513.l445D6Xh025247@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Snoop operation Date: Thu, 03 May 2007 22:13:06 -0700 From: william@bourbon.usc.edu Someone wrote: > On p. 34 of Lec. 24, what does "greater than last ACKed, > pass on" correspond to in the flowchart on the left? This corresponds to the "No" arrow out of "In-sequence". This is the case of a congestion loss. > Also, in the flowchart, if a new packet is out of sequence, > how could we immediately say that it is "congestion loss" > (the red font under the box)? I think the idea of snoop > is to abate TCP's overreaction to packet loss (some > caused by wireless environment not congestion). "Congestion loss" means that a packet is dropped by a congested router in the Internet. This is what TCP assumes when any packet is lost. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed May 2 14:58:56 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l42Lwp91006443; Wed, 2 May 2007 14:58:51 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l42LvxMn027944; Wed, 2 May 2007 14:57:59 -0700 Message-Id: <200705022157.l42LvxMn027944@bourbon.usc.edu> To: cs558l@merlot.usc.edu, cs551@merlot.usc.edu, cs530@merlot.usc.edu Subject: FWD: "Securing Peer-to-peer Systems" - CS Colloquium Talk - 3:30pm May 3 Date: Wed, 02 May 2007 14:57:59 -0700 From: william@bourbon.usc.edu Hi, Just forwarding information about a CS Colloquium talk with a topic that may be of interest to some of you. -- Bill Cheng // bill.cheng@usc.edu -------- Original Message -------- Subject: [colloq_org_stu] USC CS Colloquium Talk Reminder: May3 Date: Wed, 02 May 2007 14:00:02 -0700 From: Sumit Rangwala To: CS Colloquia Student Organizers Securing Peer-to-peer Systems Prof. Yongdae Kim University of Minnesota May 3, 2007 3:30 - 4:50 pm SSL 150 Host: Prof. Ramesh Govindan Abstract: The emergence of peer-to-peer (P2P) systems and applications signifies a revolutionary paradigm shift in building large-scale distributed systems over the Internet. This talk focuses on cooperative peer-to-peer systems, in which peer nodes are more close-knit, forming a "community of common interest" with shared goals and mutual benefits. Since the value and utility of a cooperative P2P system lie in the service it offers to its users, ensuring its correct and efficient operation despite the existence of potentially untrustworthy nodes is of utmost importance. Several infrastructure-level, currently-deployed cooperative P2P systems are vulnerable to various attacks by participating nodes that can impede the service provided by the system. Other systems are vulnerable to misuse or abuse of the services by a few peers. To prevent these attacks, we have developed two core mechanisms that can be used to build trustworthy, accountable and secure cooperative P2P systems. In the first part of the talk, I present a secure overlay routing mechanism, called Myrmic, that guarantees efficient and robust delivery of message even when significant fraction (say 30%) of nodes are compromised. Our experimental and simulation results confirm efficiency and robustness of Myrmic. The second part of the talk introduces how to "emulate" central entity on P2P systems, which is robust against collaboration attacks. Such central entity can be used for enforcing security policy as in traditional client-server systems. We also show how to enforce quota (or guarantee fairness) in a P2P file archiving system using this emulated central entities. Bio: Prof. Yongdae Kim has been working on various projects in data and communication security. During his PhD study, he designed group key agreement protocols, TGDH and STR protocol, which are integrated with Secure Spread. He received NSF career award on storage security and McKnight Land-Grant Professorship Award from University of Minnesota in 2005. His research interests include security issues for distributed systems such as P2P systems, storage systems, sensor and ad hoc networks. ------------------------------------------------------------------- Schedule for upcoming seminar available at http://enl.usc.edu/colloq _______________________________________________ colloq_org_stu mailing list colloq_org_stu@enl.usc.edu http://enl.usc.edu/mailman/listinfo/colloq_org_stu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed May 2 12:29:43 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l42JThxH026867 for ; Wed, 2 May 2007 12:29:43 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l42JSqvE026419 for ; Wed, 2 May 2007 12:28:52 -0700 Message-Id: <200705021928.l42JSqvE026419@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: slice 33 lecture 28 Date: Wed, 02 May 2007 12:28:52 -0700 From: william@bourbon.usc.edu Someone wrote: > For the RPF checking, How can I know if there exist the unicast > direction to the sending host? checking the routing table about > the unicast direction? Exactly! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed May 2 09:59:01 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l42Gx1xT014948 for ; Wed, 2 May 2007 09:59:01 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l42GwAhg024521 for ; Wed, 2 May 2007 09:58:10 -0700 Message-Id: <200705021658.l42GwAhg024521@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 16 slide 21 Date: Wed, 02 May 2007 09:58:10 -0700 From: william@bourbon.usc.edu Someone wrote: > I wanted to say that what we have in the slide indicated > as *knee* should be the *cliff*. thats it. If you are talking about only slide 21 of lecture 16, you are correct and I have just fixed the slides. But if you are talking about slide 50 of lecture 12 and slide 4 of lecture 13, then I would still disagree. According to what you have mentioned (from page 159 paragraph 2 of [Ramakrishnan90a]: We call the point at which throughput falls off rapidly before the network reaches congestion collapse as the cliff. This matches with what we have on slide 4 of lecture 13: http://merlot.usc.edu/cs551-s07/lectures/lecture13/slides-4.html So, we are consistent with the paper. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Tuesday, May 1, 2007 9:58 pm Subject: Re: lecture 16 slide 21 To: cs551@merlot.usc.edu > Someone wrote: > > > Its kind of confusing for me what u replied back in the last > > mail. Actually its clearly indicated in the research paper by > > Ramakrishnan page no 159 paragraph 2 > > > > "We call the point at which throughput falls > > off rapidly before the network reaches congestion collapse as the > > cliff. This is > > also the point at which the response time approaches infinity. > > The purpose of a > > congestion control scheme [2, 14] is to detect the fact that the > > network has > > reached the cliff, resulting in packet losses, and to reduce the > > load so that the > > network can recover to an uncongested state. " > > So far so good... Like you said, this is about the "cliff". > > > this clearly means that they also mean the cliff and knee as the > > same thing as in congestion control mechanism > > I don't know why you said that! If you read the next two > sentenses, it defines the "knee" to be at a different point > (same as what we showed in our slides). > > > So i guess there should be no reason for multiple understanding > > of knee and cliff > > I would disagree. > -- > Bill Cheng // bill.cheng@usc.edu > > > > > ----- Original Message ----- > From: william@bourbon.usc.edu > Date: Tuesday, May 1, 2007 1:44 pm > Subject: Re: lecture 16 slide 21 > To: cs551@merlot.usc.edu > > > Hi, > > > > Just want to mention that not all people agree on the > > terminology used in [Ramakrishnan90a]. In [Ramakrishnan90a], > > the authors used the term, "knee of the delay curve" to mean > > what others would call the "knee of the throughput curve". > > > > Others would argue that the "knee of a curve" is the part > > that "bend the most" (i.e., smallest radius of curvature). > > If you look at the delay curve, it bends the most at the > > point that corresponds to the "cliff of the throughput curve". > > -- > > Bill Cheng // bill.cheng@usc.edu > > > > > > > > -----Original Message----- > > Date: Mon, 30 Apr 2007 22:00:12 -0700 > > From: william@bourbon.usc.edu > > To: cs551@merlot.usc.edu > > Subject: Re: lecture 16 slide 21 > > > > Someone wrote: > > > > > In lecture 16 slide no 21(power and load) I guess in the > graph > > the Knee line is shown wrong and it should be the > first line.. > > > > You are correct. There are other slides that show the knee > > at the right place: > > > > slide 50 of lecture 12 > > slide 4 of lecture 13 > > > > Thanks for catching this one. > > -- > > Bill Cheng // bill.cheng@usc.edu > > Return-Path: william@bourbon.usc.edu Delivery-Date: Tue May 1 21:58:11 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l424wB2Z022916 for ; Tue, 1 May 2007 21:58:11 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l424vMns017959 for ; Tue, 1 May 2007 21:57:22 -0700 Message-Id: <200705020457.l424vMns017959@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 16 slide 21 Date: Tue, 01 May 2007 21:57:22 -0700 From: william@bourbon.usc.edu Someone wrote: > Its kind of confusing for me what u replied back in the last > mail. Actually its clearly indicated in the research paper by > Ramakrishnan page no 159 paragraph 2 > > "We call the point at which throughput falls > off rapidly before the network reaches congestion collapse as the > cliff. This is > also the point at which the response time approaches infinity. > The purpose of a > congestion control scheme [2, 14] is to detect the fact that the > network has > reached the cliff, resulting in packet losses, and to reduce the > load so that the > network can recover to an uncongested state. " So far so good... Like you said, this is about the "cliff". > this clearly means that they also mean the cliff and knee as the > same thing as in congestion control mechanism I don't know why you said that! If you read the next two sentenses, it defines the "knee" to be at a different point (same as what we showed in our slides). > So i guess there should be no reason for multiple understanding > of knee and cliff I would disagree. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Tuesday, May 1, 2007 1:44 pm Subject: Re: lecture 16 slide 21 To: cs551@merlot.usc.edu > Hi, > > Just want to mention that not all people agree on the > terminology used in [Ramakrishnan90a]. In [Ramakrishnan90a], > the authors used the term, "knee of the delay curve" to mean > what others would call the "knee of the throughput curve". > > Others would argue that the "knee of a curve" is the part > that "bend the most" (i.e., smallest radius of curvature). > If you look at the delay curve, it bends the most at the > point that corresponds to the "cliff of the throughput curve". > -- > Bill Cheng // bill.cheng@usc.edu > > > > -----Original Message----- > Date: Mon, 30 Apr 2007 22:00:12 -0700 > From: william@bourbon.usc.edu > To: cs551@merlot.usc.edu > Subject: Re: lecture 16 slide 21 > > Someone wrote: > > > In lecture 16 slide no 21(power and load) I guess in the graph > > the Knee line is shown wrong and it should be the first line.. > > You are correct. There are other slides that show the knee > at the right place: > > slide 50 of lecture 12 > slide 4 of lecture 13 > > Thanks for catching this one. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Tue May 1 13:44:00 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l41Ki09A016787 for ; Tue, 1 May 2007 13:44:00 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l41KhCrg010717 for ; Tue, 1 May 2007 13:43:12 -0700 Message-Id: <200705012043.l41KhCrg010717@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 16 slide 21 Date: Tue, 01 May 2007 13:43:12 -0700 From: william@bourbon.usc.edu Hi, Just want to mention that not all people agree on the terminology used in [Ramakrishnan90a]. In [Ramakrishnan90a], the authors used the term, "knee of the delay curve" to mean what others would call the "knee of the throughput curve". Others would argue that the "knee of a curve" is the part that "bend the most" (i.e., smallest radius of curvature). If you look at the delay curve, it bends the most at the point that corresponds to the "cliff of the throughput curve". -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Mon, 30 Apr 2007 22:00:12 -0700 From: william@bourbon.usc.edu To: cs551@merlot.usc.edu Subject: Re: lecture 16 slide 21 Someone wrote: > In lecture 16 slide no 21(power and load) I guess in the graph > the Knee line is shown wrong and it should be the first line.. You are correct. There are other slides that show the knee at the right place: slide 50 of lecture 12 slide 4 of lecture 13 Thanks for catching this one. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 30 22:00:57 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l4150vBC007038 for ; Mon, 30 Apr 2007 22:00:57 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l4150CcA000992 for ; Mon, 30 Apr 2007 22:00:12 -0700 Message-Id: <200705010500.l4150CcA000992@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: lecture 16 slide 21 Date: Mon, 30 Apr 2007 22:00:12 -0700 From: william@bourbon.usc.edu Someone wrote: > In lecture 16 slide no 21(power and load) I guess in the graph > the Knee line is shown wrong and it should be the first line.. You are correct. There are other slides that show the knee at the right place: slide 50 of lecture 12 slide 4 of lecture 13 Thanks for catching this one. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 30 21:53:39 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l414rd7L006269 for ; Mon, 30 Apr 2007 21:53:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l414qrGN000926 for ; Mon, 30 Apr 2007 21:52:53 -0700 Message-Id: <200705010452.l414qrGN000926@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Re B's Routing Table Date: Mon, 30 Apr 2007 21:52:53 -0700 From: william@bourbon.usc.edu Someone wrote: > I would like to know whether the slide no 23 in Freenet the 2 > routing tables named B's Routing Table are correct or whether > therre is a typo and it is A's routin table? Oops! You are correct. From slides 21 through 23, it should say A's Routing Table right on top of node A. Thanks for catching it. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 30 15:04:31 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3UM4VRl006873 for ; Mon, 30 Apr 2007 15:04:31 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3UM3kCx028398 for ; Mon, 30 Apr 2007 15:03:46 -0700 Message-Id: <200704302203.l3UM3kCx028398@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Hello Date: Mon, 30 Apr 2007 15:03:46 -0700 From: william@bourbon.usc.edu Someone wrote: > I don't know whether it is appropriate to ask or no ? It's no problem for me. > I was just wondering how would be your grading scheme. > How do you grade the average and people above average !! Well, I plug all your scores into the equation shown at: http://merlot.usc.edu/cs551-s07/description.html#grading and then I plot all the outcomes on a curve. The average score (and scores around it) will be a B+. That much is guaranteed. The following is my approximate algorithm for assigning the rest of the grades. (The word "suitable" below is not defined very precisely. It kind of means "around the right place".) Starting from the average, I look for a suitable gap upward. Above that gap, it will be A-. Then I continue to look for another suitable gap upward. Above this gap, it will be A. Similarly, starting from the average, I look for a suitable gap downward. Below this gap, it will be B. Then I continue to look for another suitable gap downward. Below this gap, it will be B-. Then I continue to look for suitable gaps downward and assign lower grades, and so on. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 29 20:30:11 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3U3UBR8014121 for ; Sun, 29 Apr 2007 20:30:11 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3U3TUwe015239 for ; Sun, 29 Apr 2007 20:29:30 -0700 Message-Id: <200704300329.l3U3TUwe015239@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: HW3 Date: Sun, 29 Apr 2007 20:29:30 -0700 From: william@bourbon.usc.edu Someone wrote: > Have you collected HW3? So far I have not received > email from Bistro about my submission. Oops! Forgot all about it! I'll do it soon. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 29 18:50:55 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3U1otiW006345 for ; Sun, 29 Apr 2007 18:50:55 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3U1oEnb013994 for ; Sun, 29 Apr 2007 18:50:14 -0700 Message-Id: <200704300150.l3U1oEnb013994@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: First-hop router in DiffServ Date: Sun, 29 Apr 2007 18:50:14 -0700 From: william@bourbon.usc.edu Someone wrote: > On p. 12 of the 2-bit DiffServ slides, we talk about the > way first-hop routers treat customer's traffic. > > It seems that out-of-profile premium packets are > penalized more harshly (delayed or dropped) than > out-of-profile assured packets (just mark them > as out-of-profile). The same can be said with > what happens in border router (drop extra P packet > but sends, not drop, extra A packet to forwarding engine, > as on p. 14). > > Why so? Should premium service get "better" > treatment than assured service? I assume you meant "worse" instead of "better"... The reason is that premium service do get treated better in the network (has high priority over others -- see bottom of slide 15 of lecture 22). If you downgrade permium packets when they arrive without tokens, then you will end up with packet received out of order, and that's really bad. When an assured packet get downgraded, it stays in the same queue but can get dropped with higher probability (but no packet out of order). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 28 13:26:25 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3SKQPtq003690 for ; Sat, 28 Apr 2007 13:26:25 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3SKPmCc005383 for ; Sat, 28 Apr 2007 13:25:48 -0700 Message-Id: <200704282025.l3SKPmCc005383@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: End-game mode for BitTorrent Date: Sat, 28 Apr 2007 13:25:48 -0700 From: william@bourbon.usc.edu Someone wrote: > We have the following on p. 27 of the following slide: > > http://merlot.usc.edu/cs551-s07/lectures/lecture20/slides.pdf > > What does it mean to "send all sub-pieces request to all > peers after all data sub-pieces request are sent?" > > Specifically, is there any difference between the two "sub-pieces" > mentioned here? I asked the person who created this slide and after some investigation, it seemed that I may have explained things incorrectly regarding the "end-game" mode. In class, I explained that the "end-game" mode is used when there is only one piece of data that a node does not have, so it sends requests for all the sub-pieces of this piece to all peers. This may be a wrong explanation. It looks like the "end-game" mode applies to *every piece* of data. Typically, a piece is 256KB and it is divided into 16 sub-pieces (16KB each). For a piece that has missing sub-pieces, it may send a request for a sub-piece to one peer and send a request for another sub-piece to a different peer. Internally, this node keeps track of what sub-piece requests have been sent. When *all the sub-piece requests have been sent*, it enters the "end-game" mode for this piece and *sends all sub-pieces request to all peers*. I think this explanation is better than the one given inclass. (I haven't read the code, so I don't know for sure if this interpretation is 100% correct.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 28 08:19:23 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3SFJNGY011756 for ; Sat, 28 Apr 2007 08:19:23 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3SFIlOJ002978 for ; Sat, 28 Apr 2007 08:18:47 -0700 Message-Id: <200704281518.l3SFIlOJ002978@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: traffic shaping Date: Sat, 28 Apr 2007 08:18:47 -0700 From: william@bourbon.usc.edu Someone wrote: > When a traffic is shaped (delayed from entering a network), > exactly where in the boundary the event occurs? Is it > inside the boundary router (i.e. no tokens, queue packet up) or > its input port (that would seem like "drop" the packet)? If you are asking a generic question, then anywhere is possible. Wherever you want the traffic to be shaped, you can install a token bucket filter. Or are you asking just about an input port? If you put a token bucket filter right before an input, that would work because a token bucket filter can has its own buffer. (A token bucket filter is a mathematical abstraction, it can be implemented any way you want.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 28 08:13:43 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3SFDhe2011288 for ; Sat, 28 Apr 2007 08:13:43 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3SFD7UV002908 for ; Sat, 28 Apr 2007 08:13:07 -0700 Message-Id: <200704281513.l3SFD7UV002908@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: End-game mode for BitTorrent Date: Sat, 28 Apr 2007 08:13:07 -0700 From: william@bourbon.usc.edu Someone wrote: > Does "based on the uploading rate" here imply that > the leeches with lower downstream traffic will > be favored by seeder? Normally, whom to unchoke depends on the "download rate" of peers. When a node becomes a seeder, no peer will be downloading to it. So, looking at downloadingn rate is useless. At this point, looking at "uploading rate" makes sense because for a seeder, it is good to get its data out as fast as possible. So, the higher the peer's upload rate the better. Therefore, the policy is to unchoke peers with the highest upload rates. (I guess an alternative is to contact the tracker and see which piece is the rarest in the torrent and use that as a priority to decide which peer to upload to. But this requires contactingn the tracker and that's not desirable.) -- Bill Cheng // bill.cheng@usc.edu On 4/27/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > On p. 29 of lecture 20, it reads that seeders will > > unchoke peers "based on the uploading rate", then > > "preferring peers which no one else happens to be > > uploading to them at the moment". > > > > Are the two statements counter to each other? > > I don't think so... I think the way to interpret the > last bullet is: > > as a result, this would prefer peers which no one else > happens to be uploading to them at that moment > > Does this make more sense now? Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 27 23:19:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3S6JY3S001000 for ; Fri, 27 Apr 2007 23:19:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3S6Ixdg026858 for ; Fri, 27 Apr 2007 23:18:59 -0700 Message-Id: <200704280618.l3S6Ixdg026858@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: End-game mode for BitTorrent Date: Fri, 27 Apr 2007 23:18:59 -0700 From: william@bourbon.usc.edu Someone wrote: > On p. 29 of lecture 20, it reads that seeders will > unchoke peers "based on the uploading rate", then > "preferring peers which no one else happens to be > uploading to them at the moment". > > Are the two statements counter to each other? I don't think so... I think the way to interpret the last bullet is: as a result, this would prefer peers which no one else happens to be uploading to them at that moment Does this make more sense now? > And is the second statement similar to "optimistic unchoking" > on the next page? I don't think so either. The word "optimistic" here is probably not the best name for it. In this case, it just means "regardless of download rate". So, it's giving all other connections a chance by unchoking them in a round-robin fashion. (Is this really "optimistic"? Can you find a better word for it?) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 27 07:44:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3REiITV024618 for ; Fri, 27 Apr 2007 07:44:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3REhjGU017347 for ; Fri, 27 Apr 2007 07:43:45 -0700 Message-Id: <200704271443.l3REhjGU017347@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Steady-state queuing Date: Fri, 27 Apr 2007 07:43:45 -0700 From: william@bourbon.usc.edu Someone wrote: > On p. 31 of the DEC-bit slide, we have: > > Full queues problem: > > Drop tail maintains full or nearly-full queues > during congestion; but queue limits should reflect > the size of bursts we want to absorb, not steady-state > queuing. > > I do not quite understand "steady-state queuing", > can you elaborate? Steady-state means "long-term equilibrium". (Of course, in networking, "long-term" could just mean seconds or minutes.) Queueing means queue built up, which implies there is a congestion of some sort (not congestion collapse). So, "steady-state queueing" can be think of as related to "long-term congestion". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 25 18:24:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3Q1OXYh009601 for ; Wed, 25 Apr 2007 18:24:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3Q1O6UM027209 for ; Wed, 25 Apr 2007 18:24:06 -0700 Message-Id: <200704260124.l3Q1O6UM027209@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Office hours Date: Wed, 25 Apr 2007 18:24:06 -0700 From: william@bourbon.usc.edu Someone wrote: > Will you still hold office hours during the study week? No! But you are welcome to make appointments! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 25 10:28:16 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3PHSGKO004418 for ; Wed, 25 Apr 2007 10:28:16 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3PHRpSI022182 for ; Wed, 25 Apr 2007 10:27:51 -0700 Message-Id: <200704251727.l3PHRpSI022182@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Final exam topics Date: Wed, 25 Apr 2007 10:27:51 -0700 From: william@bourbon.usc.edu Someone wrote: > Although we have the syllabus and slides, can you also > post topics for the final exam as you did for the midterm? I was going to wait till the last lecture is over. But since I'm sure that we will finish covering Single Source Multicast and Scalable Reliable Multicast, I've posted a summary of topics on the class web page. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 22:20:44 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3O5KiaB025606 for ; Mon, 23 Apr 2007 22:20:44 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3O5KOlP028607 for ; Mon, 23 Apr 2007 22:20:24 -0700 Message-Id: <200704240520.l3O5KOlP028607@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: midterm questions Date: Mon, 23 Apr 2007 22:20:24 -0700 From: william@bourbon.usc.edu Someone wrote: > Just a friendly reminder, is your message intended for 530 class? > I am not quite sure since we also covered Bistro in last lecture, though. Ooops! Sorry! It was for CS 530. -- Bill Cheng // bill.cheng@usc.edu On 4/23/07, william@bourbon.usc.edu wrote: > Someone wrote: > > [ I think the subject line above should be "Exam #2 Questions". ] > > > In Intrusion Detection topic, the last bullet is "response", in > > the lecture you skipped the slides for "phases of response" will > > those be in the final? > > Since it's part of the textbook, it is included in the topics > for Exam #2. > > > Also, on the "Upload" topic, will the questions be about bistro? > > how it works, goals, and how bistro "solves" the deadline > > problem? > > Yes. > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 22:13:36 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3O5DarP024956 for ; Mon, 23 Apr 2007 22:13:36 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3O5DG5n028417 for ; Mon, 23 Apr 2007 22:13:16 -0700 Message-Id: <200704240513.l3O5DG5n028417@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Final Date: Mon, 23 Apr 2007 22:13:16 -0700 From: william@bourbon.usc.edu Someone wrote: > Can you confirm with us that the material to be tested in the > final *starts* from this date: > > (Mar 1, 2007) - DEC-bit [Ramakrishnan90a], RED [Floyd93a] Correct. -- Bill Cheng // bill.cheng@usc.edu On 4/23/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > Is the final comprehensive or just what we learnt after the mid term? > > Only covers what was covered after the midterm. > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 22:10:15 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3O5AFV3024782 for ; Mon, 23 Apr 2007 22:10:15 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3O59tiS028322 for ; Mon, 23 Apr 2007 22:09:55 -0700 Message-Id: <200704240509.l3O59tiS028322@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: midterm questions Date: Mon, 23 Apr 2007 22:09:55 -0700 From: william@bourbon.usc.edu Someone wrote: [ I think the subject line above should be "Exam #2 Questions". ] > In Intrusion Detection topic, the last bullet is "response", in > the lecture you skipped the slides for "phases of response" will > those be in the final? Since it's part of the textbook, it is included in the topics for Exam #2. > Also, on the "Upload" topic, will the questions be about bistro? > how it works, goals, and how bistro "solves" the deadline > problem? Yes. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 21:16:37 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3O4GbLl020444 for ; Mon, 23 Apr 2007 21:16:37 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3O4GHj1026951 for ; Mon, 23 Apr 2007 21:16:17 -0700 Message-Id: <200704240416.l3O4GHj1026951@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Final Date: Mon, 23 Apr 2007 21:16:17 -0700 From: william@bourbon.usc.edu Someone wrote: > Is the final comprehensive or just what we learnt after the mid term? Only covers what was covered after the midterm. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 10:29:28 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3NHTOQR001393; Mon, 23 Apr 2007 10:29:24 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3NHT62P018299; Mon, 23 Apr 2007 10:29:06 -0700 Message-Id: <200704231729.l3NHT62P018299@bourbon.usc.edu> To: cs530@merlot.usc.edu, cs551@merlot.usc.edu, cs558l@merlot.usc.edu Subject: Re: CSCI 530 relevant news article . . . Date: Mon, 23 Apr 2007 10:29:06 -0700 From: william@bourbon.usc.edu Thanks for sending it! I guess sonner or later, this is going to happen since IRS would not accept better solution to their scalability problem. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Mon, 23 Apr 2007 06:23:48 -0700 To: william@bourbon.usc.edu Subject: CSCI 530 relevant news article . . . Per this article, it sounds like Intuit should have been using Bistro ... http://www.nytimes.com/2007/04/23/technology/23intuit.html ;) Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 23 10:10:46 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3NHAkXP032435 for ; Mon, 23 Apr 2007 10:10:46 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3NHASJ7017976 for ; Mon, 23 Apr 2007 10:10:28 -0700 Message-Id: <200704231710.l3NHASJ7017976@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: 404 Error Date: Mon, 23 Apr 2007 10:10:28 -0700 From: william@bourbon.usc.edu Someone wrote: > I am getting "404 File Not Found" error upon accessing the > lecture slides for 19-Apr. Please help me out. Ooops! Fixed. Thanks! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 22 20:38:26 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3N3cQ7E025970 for ; Sun, 22 Apr 2007 20:38:26 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3N3cAbu009516 for ; Sun, 22 Apr 2007 20:38:10 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l3N3cAt8009515 for cs551@merlot; Sun, 22 Apr 2007 20:38:10 -0700 Date: Sun, 22 Apr 2007 20:38:10 -0700 From: william@bourbon.usc.edu Message-Id: <200704230338.l3N3cAt8009515@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: final project part (2) is due tonight... Hi, Please do *not* submit late! The late policy will be followed *strictly* to be fair to all. Make multiple submissions if you'd like. Please check the server time to synchronize it with your watch by visiting the following link: http://merlot.usc.edu:9996/bistro/myip.html Please also remember that you have 24 hours to make 3 free lines of changes after the deadline. You should take advantage of that. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 22 19:52:17 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2qHrS022329 for ; Sun, 22 Apr 2007 19:52:17 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2q1Wo008992 for ; Sun, 22 Apr 2007 19:52:01 -0700 Message-Id: <200704230252.l3N2q1Wo008992@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Really Weird Behavior Date: Sun, 22 Apr 2007 19:52:01 -0700 From: william@bourbon.usc.edu Someone wrote: > For the last 2-3 hours I was trying to fix a very weird bug where the > metafiles and datafiles that I transfer to other nodes would be sometimes > totally corrupted (with truly weird outputs such as other people's names** > and the word "borbon" in it) -OR- only having some difference in the last > line... > > And finally, I changed how I allocate my buffer that holds the metafile > and datafile from 1. to 2. like this: > > 1. char* msg_buf = (char*)malloc(msg_buf_sz); > 2. char msg_buf[msg_buf_sz]; > > and then, automagically everything started working flawlessly. What gives? > > So now my question is, what is the difference between those 2 lines and > what exactly happened when I was using line 1? > > ** - so then of course I check who's on nunki with "top" but the person's > name didn't match anyone online... :| Hmm... I didn't even know that (2) above is allowed! Sounded to me like you have an uninitialized memory bug and you are picking up random data on the stack. Another possibility is that it's a memory corruption bug and you are picking up data from somewhere else. It should have nothing to do with other people on the system since your virtual memory space is separate from others' (unless you are using shsared memory, which I assume you are not). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 22 19:47:32 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2lWVU021850 for ; Sun, 22 Apr 2007 19:47:32 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2lGh5008898 for ; Sun, 22 Apr 2007 19:47:16 -0700 Message-Id: <200704230247.l3N2lGh5008898@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Logging in single beacon Date: Sun, 22 Apr 2007 19:47:16 -0700 From: william@bourbon.usc.edu Someone wrote: > Just to make sure, in a single beacon scenario, > is it OK to have an empty log? Since you are not sending, forwarding, or receiving anything, it's okay to have an empty log. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 22 19:44:53 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2irOO021680 for ; Sun, 22 Apr 2007 19:44:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3N2ibMN008843 for ; Sun, 22 Apr 2007 19:44:37 -0700 Message-Id: <200704230244.l3N2ibMN008843@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: kwrd,name,sha1 index files format Date: Sun, 22 Apr 2007 19:44:37 -0700 From: william@bourbon.usc.edu Someone wrote: > Is there a specific format for writing names,keywords and sha > values into name_index,kwrd_index and sha1_index files? No. It's your design decision. If you haven't implemented this yet, since it's only a few hours away from the submission deadline, I think the best way to spend your time is to review the grading guidelines and get some bugs out and try to get as many points as possible instead of implement something new and risk breaking something that's already working. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 22 09:29:46 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3MGTkno005183 for ; Sun, 22 Apr 2007 09:29:46 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3MGTVg2002831 for ; Sun, 22 Apr 2007 09:29:31 -0700 Message-Id: <200704221629.l3MGTVg2002831@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About delete Date: Sun, 22 Apr 2007 09:29:31 -0700 From: william@bourbon.usc.edu Someone wrote: > On the node receiving DELETE msg, is computing the digital > signature of the FileSpec and comparing it against the received > filespec the same as using openssl to verify delete? If this node uses its own private key to "compute a digital signature", then with 100% certainty it will not "match" against the received FileSpec. What's in a DELETE message is a *digitally signed FileSpec*. Then only thing you can do with such a thing is to *verify the digital signature* (which requires the public key of the signer, and this public key is inside the certificate of the signer). Therefore, when a node receives a DELETE message, it should just verify the digital signature. Please see: http://merlot.usc.edu/cs551-s07/projects/final.html#cert regarding how to verify a digital signature. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, April 21, 2007 10:46 pm Subject: Re: About delete To: cs551@merlot.usc.edu > Someone wrote: > > > You said : > > START QUOTE================================= > > You should see that the following 3 lines are visible to your > > program: > > > > FileName=... > > SHA1=... > > Nonce=... > > > > So, you can extract the FileName, SHA1, and Nonce values from > > a signed File Spec *before* to try to verify any digital > > signatures. Then you can use, say, the SHA1 index structure > > to find all files with the same SHA1 files, open their > > metadata files and find a subset that has the same Noce > > (hopefullly, there is only one such file since you should not > > have duplicates), then apply the corresponding certificate > > file to verify digital signature. > > > > ===================END QUOTE > > > > We still have to match the filename if all the above is > > found to be matching? > > You need to find the file number in your mini-filesystem > that has exactly the same FileName, SHA1, and Nonce. > -- > Bill Cheng // bill.cheng@usc.edu > > > > > ----- Original Message ----- > From: william@bourbon.usc.edu > Date: Tuesday, April 17, 2007 7:36 pm > Subject: Re: About delete > To: cs551@merlot.usc.edu > > > Someone wrote: > > > > > I am not clear how the delete works? > > > When a node gets a delete packet it verifies it with all > the > > .pem files > > > he has in his home directory and then matches the > > contents(file name, > > > sha1, nonce) provided in the delete packet > > > nad then deletes the file.. > > > is it the right flow isn't it very inefficient or am I > > missing something? > > > > This is inefficient. If you look at the sample signed File Spec: > > > > http://merlot.usc.edu/cs551-s07/projects/cert/signed.txt > > > > You should see that the following 3 lines are visible to your > > program: > > > > FileName=... > > SHA1=... > > Nonce=... > > > > So, you can extract the FileName, SHA1, and Nonce values from > > a signed File Spec *before* to try to verify any digital > > signatures. Then you can use, say, the SHA1 index structure > > to find all files with the same SHA1 files, open their > > metadata files and find a subset that has the same Noce > > (hopefullly, there is only one such file since you should not > > have duplicates), then apply the corresponding certificate > > file to verify digital signature. > > -- > > Bill Cheng // bill.cheng@usc.edu > > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 22:46:41 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5kf6e025890 for ; Sat, 21 Apr 2007 22:46:41 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5kS0u004130 for ; Sat, 21 Apr 2007 22:46:28 -0700 Message-Id: <200704220546.l3M5kS0u004130@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: regarding delete Date: Sat, 21 Apr 2007 22:46:28 -0700 From: william@bourbon.usc.edu Someone wrote: > I just wanted to ask where we're supposed to store the private an public > keys for a particular node. I mean should they be in the current working > directory or in the node directory? Somewhere in HomeDir. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 22:45:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5jYRb025727 for ; Sat, 21 Apr 2007 22:45:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5jLgU004087 for ; Sat, 21 Apr 2007 22:45:21 -0700 Message-Id: <200704220545.l3M5jLgU004087@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About delete Date: Sat, 21 Apr 2007 22:45:21 -0700 From: william@bourbon.usc.edu Someone wrote: > You said : > START QUOTE================================= > You should see that the following 3 lines are visible to your > program: > > FileName=... > SHA1=... > Nonce=... > > So, you can extract the FileName, SHA1, and Nonce values from > a signed File Spec *before* to try to verify any digital > signatures. Then you can use, say, the SHA1 index structure > to find all files with the same SHA1 files, open their > metadata files and find a subset that has the same Noce > (hopefullly, there is only one such file since you should not > have duplicates), then apply the corresponding certificate > file to verify digital signature. > > ===================END QUOTE > > We still have to match the filename if all the above is > found to be matching? You need to find the file number in your mini-filesystem that has exactly the same FileName, SHA1, and Nonce. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Tuesday, April 17, 2007 7:36 pm Subject: Re: About delete To: cs551@merlot.usc.edu > Someone wrote: > > > I am not clear how the delete works? > > When a node gets a delete packet it verifies it with all the > .pem files > > he has in his home directory and then matches the > contents(file name, > > sha1, nonce) provided in the delete packet > > nad then deletes the file.. > > is it the right flow isn't it very inefficient or am I > missing something? > > This is inefficient. If you look at the sample signed File Spec: > > http://merlot.usc.edu/cs551-s07/projects/cert/signed.txt > > You should see that the following 3 lines are visible to your > program: > > FileName=... > SHA1=... > Nonce=... > > So, you can extract the FileName, SHA1, and Nonce values from > a signed File Spec *before* to try to verify any digital > signatures. Then you can use, say, the SHA1 index structure > to find all files with the same SHA1 files, open their > metadata files and find a subset that has the same Noce > (hopefullly, there is only one such file since you should not > have duplicates), then apply the corresponding certificate > file to verify digital signature. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 22:15:35 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5FZfR023189 for ; Sat, 21 Apr 2007 22:15:35 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5FMi8003849 for ; Sat, 21 Apr 2007 22:15:22 -0700 Message-Id: <200704220515.l3M5FMi8003849@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on DELETE Date: Sat, 21 Apr 2007 22:15:22 -0700 From: william@bourbon.usc.edu Someone wrote: > No I was asking it the user can use the DELETE command in our > assignment to delete a file locally if it didnt do the > corresponding STORE but got the file using GET. It should verify the digital signature and this should fail. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, April 21, 2007 8:41 pm Subject: Re: Question on DELETE To: cs551@merlot.usc.edu > Someone wrote: > > > Since only the user who created a file using store can delete it > > from the network. So a user who GOT a file cant delete it even > > from his node's permanent storage, right? and its not being asked > > for in this assignment? > > If this user has write privilege in HomeDir, he/she can > delete the file from the file system by doing "rm" and there > is no way to prevent this user from doing this. Is that what > you were asking? > > The only way specified in the assignment to delete a file is > via the DELETE message and your node should verify digital > signature before deleting a file. > -- > Bill Cheng // bill.cheng@usc.edu > > > > ----- Original Message ----- > From: william@bourbon.usc.edu > Date: Friday, April 13, 2007 3:19 pm > Subject: Re: Question on DELETE > To: cs551@merlot.usc.edu > > > Someone wrote: > > > > > Say two nodes have the same copy of a file (filename, > nonce, and > > > sha1), and one of them try to 'delete'. It first tries to > > delete the local > > > copy, however the signature verification fails. Should it > stop > > or continue > > > flooding out Delete to other nodes? > > > > Yes. I guess it's possible that the local copy is somehow > corrupted. > -- > > Bill Cheng // bill.cheng@usc.edu > > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 22:14:05 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5E59N023167 for ; Sat, 21 Apr 2007 22:14:05 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M5DrIG003815 for ; Sat, 21 Apr 2007 22:13:53 -0700 Message-Id: <200704220513.l3M5DrIG003815@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS 551 - Final Project - Grading Guidelines - Test Case (B) (1) (a) Date: Sat, 21 Apr 2007 22:13:53 -0700 From: william@bourbon.usc.edu Someone wrote: > For the following part of the test case : > > in 6th window (node *08), type: > (+1 points) > search filename=earc.jpg > should get 5 responses, > look for the FileID that matches the FileID at > node *02, let's call this x, then type: > (+1 points) > get [x] > type "status files 5 05.out", look at "05.out", > all nodes should have "earc.jpg" and each > node should only have one copy of it > > > > How will the grader know which FileID corresponds to node *02? > Do we need to log the FileID -> Node ## mapping somewhere? You must be looking at an old grading guidelines. Please see my message with timestamp "Tue 17 Apr 20:08". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 20:41:29 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3fTBV015729 for ; Sat, 21 Apr 2007 20:41:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3fG9O002942 for ; Sat, 21 Apr 2007 20:41:16 -0700 Message-Id: <200704220341.l3M3fG9O002942@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on DELETE Date: Sat, 21 Apr 2007 20:41:16 -0700 From: william@bourbon.usc.edu Someone wrote: > Since only the user who created a file using store can delete it > from the network. So a user who GOT a file cant delete it even > from his node's permanent storage, right? and its not being asked > for in this assignment? If this user has write privilege in HomeDir, he/she can delete the file from the file system by doing "rm" and there is no way to prevent this user from doing this. Is that what you were asking? The only way specified in the assignment to delete a file is via the DELETE message and your node should verify digital signature before deleting a file. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Friday, April 13, 2007 3:19 pm Subject: Re: Question on DELETE To: cs551@merlot.usc.edu > Someone wrote: > > > Say two nodes have the same copy of a file (filename, nonce, and > > sha1), and one of them try to 'delete'. It first tries to > delete the local > > copy, however the signature verification fails. Should it stop > or continue > > flooding out Delete to other nodes? > > Yes. I guess it's possible that the local copy is somehow corrupted. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 20:38:12 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3cCRd015548 for ; Sat, 21 Apr 2007 20:38:12 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3c0qN002894 for ; Sat, 21 Apr 2007 20:38:00 -0700 Message-Id: <200704220338.l3M3c0qN002894@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on DELETE Date: Sat, 21 Apr 2007 20:38:00 -0700 From: william@bourbon.usc.edu Someone wrote: > When deleting the local copy we only remove it from the > home_dir/files directory right? We do not keep track of the > files in the working directory, correct? Correct. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Friday, April 13, 2007 3:19 pm Subject: Re: Question on DELETE To: cs551@merlot.usc.edu > Someone wrote: > > > Say two nodes have the same copy of a file (filename, nonce, and > > sha1), and one of them try to 'delete'. It first tries to > delete the local > > copy, however the signature verification fails. Should it stop > or continue > > flooding out Delete to other nodes? > > Yes. I guess it's possible that the local copy is somehow corrupted. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 20:25:08 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3P8GO014355 for ; Sat, 21 Apr 2007 20:25:08 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3OtMY002590 for ; Sat, 21 Apr 2007 20:24:55 -0700 Message-Id: <200704220324.l3M3OtMY002590@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About LRU Date: Sat, 21 Apr 2007 20:24:55 -0700 From: william@bourbon.usc.edu Someone wrote: > LRU list is updated > At the sender(which sent the msg) node when: > originating store? > successful get? If you originate a STORE or afer a successful GET, the file you just got should be put at the end of the LRU list. If you already have the file in the GET case, then you should move the file to the end of the LRU list. > At the receiver(which sent the response) node: > receiving store? > receiving get? Again, when you STORE a file, you put it at the end of the LRU list. You don't need to do anything when you get a GET. You need to move a file to the end of the LRU list if it matches a SEARCH request. > At intermediate nodes(just fwding msgs or responses): > fwding get response? Intermediate nodes do nothing (as far as I can think of). -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Thursday, April 12, 2007 11:44 am Subject: Re: About LRU To: cs551@merlot.usc.edu > Hi, > > Just want to clarify... What I thought the question was > about your index structure to keep track of LRU. That has no > limit. Certainly, the total size of the cahced data file (you > don't have to count metadata and certificate file sizes, and > it's okay if you do) has a limit which is specified by the > CacheSize key in the startup configuration file. > -- > Bill Cheng // bill.cheng@usc.edu > > > > -----Original Message----- > Date: Sat, 07 Apr 2007 16:18:43 -0700 > From: william@bourbon.usc.edu > To: cs551@merlot.usc.edu > Subject: Re: About LRU > > Someone wrote: > > > What is the size limit for cache used in implementing > > LRU? (e.g. bytes or number of files). > > There is no limit. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 20:21:21 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3LL6T014035 for ; Sat, 21 Apr 2007 20:21:21 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3L9vO002541 for ; Sat, 21 Apr 2007 20:21:09 -0700 Message-Id: <200704220321.l3M3L9vO002541@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on test guideline Date: Sat, 21 Apr 2007 20:21:09 -0700 From: william@bourbon.usc.edu Someone wrote: > In (A)(2)(a), CacheProb should be equal to 0 but seems in the files > b4-n00.ini, b4-n01.ini , b4-n02.ini, the CacheProb=1. Are these correct? Since these are fully connected beacon nodes, CacheProb shouldn't matter because everything should be one-hop away. But I guess it's possible that if you do things a certain way, you may end up sending data in two hops. Anyway, you are correct that they are inconsistent. I've changed the b4-n??.ini files. Thanks for catching these. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 20:14:44 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3Eivh013442 for ; Sat, 21 Apr 2007 20:14:44 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3M3EVrE002431 for ; Sat, 21 Apr 2007 20:14:31 -0700 Message-Id: <200704220314.l3M3EVrE002431@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj part2 Date: Sat, 21 Apr 2007 20:14:31 -0700 From: william@bourbon.usc.edu Someone wrote: > Should the given init neighbor list files be in the current > working dir or home dir They should be in HomeDir (just like in part (1)) so that if your code is working properly in part (1), you don't have to change anything. > and i thought the name of the file is > init_neighbor_list but the files you have given have different > names so should we write code such that the name is based on the > last two digits of the .ini file? The ~csci551b/public/final2/scripts/setup-r789 script is doing the above for you. This script must be run according to the grading guidelines. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 08:47:10 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3LFl9EZ023568 for ; Sat, 21 Apr 2007 08:47:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3LFkx9Y028225 for ; Sat, 21 Apr 2007 08:46:59 -0700 Message-Id: <200704211546.l3LFkx9Y028225@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS 551 - Final Project - Grading Guidelines - Test Case (B) (1) (a) Date: Sat, 21 Apr 2007 08:46:59 -0700 From: william@bourbon.usc.edu Someone wrote: > Say n02 is connected to beacon n00 (but n00 is not n02's neighbor). In the given topology, n02 is not connected to n00. > When n02 floods a message, it floods only to it neighbors ignoring the > beacons it is has connections to? I'm not sure why n02 would have a connection to n00. It should not. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Saturday, April 21, 2007 7:46 AM To: cs551@merlot.usc.edu Subject: Re: CS 551 - Final Project - Grading Guidelines - Test Case (B) (1) (a) Someone wrote: > For this test case the parameters are: CacheProb=1, StoreProb=1, > NeighborStoreProb=1. > > store earc.jpg 2 title="Heart of Glass" > > in 5th window (node *02), type: > "status files 5 04.out", look at "04.out", > (+1 points) > every node except node *08 should have this file, > node *08 should have no files > > Why will node *08 not contain any file if CacheProb=1, StoreProb=1, > NeighborStoreProb=1 and the store message is flooded to the entire network? Because the topology is: /------ n08 | | v v n00* <--> n01* ^^ ^ || | |\------ n04 <----\ | ^ | | | | \------- n03 <-- n02 So, if you do a STORE with a TTL of 2, it should not reach node *08. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 07:46:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3LEkYUK018749 for ; Sat, 21 Apr 2007 07:46:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3LEkNYZ027364 for ; Sat, 21 Apr 2007 07:46:23 -0700 Message-Id: <200704211446.l3LEkNYZ027364@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS 551 - Final Project - Grading Guidelines - Test Case (B) (1) (a) Date: Sat, 21 Apr 2007 07:46:23 -0700 From: william@bourbon.usc.edu Someone wrote: > For this test case the parameters are: CacheProb=1, StoreProb=1, > NeighborStoreProb=1. > > store earc.jpg 2 title="Heart of Glass" > > in 5th window (node *02), type: > "status files 5 04.out", look at "04.out", > (+1 points) > every node except node *08 should have this file, > node *08 should have no files > > Why will node *08 not contain any file if CacheProb=1, StoreProb=1, > NeighborStoreProb=1 and the store message is flooded to the entire network? Because the topology is: /------ n08 | | v v n00* <--> n01* ^^ ^ || | |\------ n04 <----\ | ^ | | | | \------- n03 <-- n02 So, if you do a STORE with a TTL of 2, it should not reach node *08. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 21 02:44:23 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3L9iNEs026732 for ; Sat, 21 Apr 2007 02:44:23 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3L9iD2f024245 for ; Sat, 21 Apr 2007 02:44:13 -0700 Message-Id: <200704210944.l3L9iD2f024245@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj Part 2 Date: Sat, 21 Apr 2007 02:44:13 -0700 From: william@bourbon.usc.edu Someone wrote: > I do the following > > STOREPROB parsed from .ini file > > float storeProbability = drand48() > > if(storeProbability < STOREPROB) > i cache the file > else > i dont > > Given STOREPROB = 0 > > if storeProbability <<0.1 say 0.02 it satisifies the condition > and stores the file when in actual it should not. This happens > for very samll values returned by drand48() > > I dont understand why?? My guess is that somehow your STOREPROB is not exactly zero. You should print it out and see what it is exactly. You should make it exactly zero when you read your INI file. Don't read it as a floating point number and just read it as a string. Check if the string is "0", if not, then call sscanf() or atof() to convert it into a float or double. By the way, I rarely use float in my code. Use double instead of float. With the kind of processors we have these days, extra steps has to be taken to convert doubles into floats. So, computation with doubles is actually faster than floats. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 23:02:06 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3L626wJ009125; Fri, 20 Apr 2007 23:02:06 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3L61up0022395; Fri, 20 Apr 2007 23:01:56 -0700 Message-Id: <200704210601.l3L61up0022395@bourbon.usc.edu> To: cs530@merlot.usc.edu, cs558l@merlot.usc.edu, cs551@merlot.usc.edu Cc: Hariram Yogendran Subject: Re: Your recommendation for candidates in a fast growing startup Date: Fri, 20 Apr 2007 23:01:56 -0700 From: william@bourbon.usc.edu Hi, I'm forwarding a job opportunity from a startup company, Inhance.Net. I've put the two Word attachments mentioned in the e-mail at the following places: Software Development Engineer Job Description: http://merlot.usc.edu/cs530-s07/misc/InhanceJob.doc Inhance Media Exec Summary: http://merlot.usc.edu/cs530-s07/misc/InhanceSummary.doc Please check the Job Description and make sure you have to qualifications before applying. Thanks! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Mon, 16 Apr 2007 14:46:41 -0700 From: Hariram Yogendran To: chengw@usc.edu Subject: Your recommendation for candidates in a fast growing startup Dear Professor Cheng, I am a recent USC Alumnus who graduated from the Computer Science Dept last year. I had taken my security systems course under you. After earning my MSCS, I was recruited by some fellow Alumni from the USC Schools of Business, Engineering and Music who have founded a very interesting, music-centric internet company. We have development centers in the United States, India and Canada and we are currently expanding our operations in the Los Angeles area. We are looking to partner with the Computer Science department at USC to find the best and brightest talent to join our team. We need bright, passionate engineers looking to really make an impact in a very dynamic, (well-funded) internet startup with tremendous potential. We are entering our site launch phase over the next couple of months so it is a very exciting time to get involved. We have very competitive compensation packages, including pre-IPO stock options and immense opportunities to work with an exceptional pool of talent. Our ideal candidate would have the following: * Primary interests in some of the following areas: networking hardware & software, security, web technologies, algorithms and databases. * Scalable vision. Big thinkers able to push the boundaries with great, forward thinking vision. * Natural problem solver with very strong Mathematical and Analytical skills. * Great aptitude, strong drive and a positive attitude. I have attached a job description for Software Development Engineers and our Management Bio's for your review. I'd like to talk further about possible students that you think may be of interest. This is a very special career opportunity and I want to thank you in advance for any assistance you may have in connecting us with up & coming talented Trojans! See our teaser web site at: www.inhance.net Looking forward to a long-term partnership with you! Sincerely, Hariram Yogendran IT Manager Inhance.Net 15912 Arminta St. Van Nuys, CA 91406 work. 818.455.4340 ext 223 cell. 213.210.9643 Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 19:53:25 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3L2rPNm026263 for ; Fri, 20 Apr 2007 19:53:25 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3L2rGFS019783 for ; Fri, 20 Apr 2007 19:53:16 -0700 Message-Id: <200704210253.l3L2rGFS019783@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Question on Grading guidelines Date: Fri, 20 Apr 2007 19:53:16 -0700 From: william@bourbon.usc.edu Someone wrote: > By the way, "if the node had to check for duplicate file", it wouldn't need > to wait until it sends a get request and then receives a "get response" from > a remote node and then see that it already has the same file in its > permanent area. It is possible to check this duplication before it self > since we have the meta data in the search results (SHA1, file name and > Nonce). Yes, it's possible. When you get your SEARCH responses, you need to keep track of the FileIDs and printout some file metadata to the screen. If you only keep the FileIDs around, then when you do a GET, you wouldn't know what the FileName, SHA1, and Nonce of the file. If you do keep the file metadata, then you can do the lookup you mentioned above. So, it depends on how you have implemented things. > My understanding is that the spec does not want us to do this. Instead, > just get the actual file that the user is asking for. In general, I like a more generic solution and not write code to check all kinds of special conditions. So, the result would not be as efficient as possible, but then you may end up with less bugs! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 19:39:39 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3L2ddX8025165 for ; Fri, 20 Apr 2007 19:39:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3L2dUIK019604 for ; Fri, 20 Apr 2007 19:39:30 -0700 Message-Id: <200704210239.l3L2dUIK019604@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Question on Grading guidelines Date: Fri, 20 Apr 2007 19:39:30 -0700 From: william@bourbon.usc.edu Someonew rote: | What is meant by: | | > 2. If the file comes from the other two nodes, it should check | > before attempting to store and find that it already has a copy of the same | > file, and discard the response. | | Don't we get files based on the "FileID" which is unique (or atleast | supposed to be) across the network. Correct. But FileID is not part of the file's metadata. It's simply an identifier for the sole purpose of retrieving the file using GET. | So when a user does "get 1" or "get 2" | etc... it actually means get the file with this particular FileID. So if a | user happens to do a get for a FileID that is remote, we still have to get | that remote copy for the user and not the local one. Right? Correct. But when you eventually receive the file, you will attempt to store it in the mini-filesystem. At this time, you should noticed that you already have a file that is identical (with the same FileName, SHA1, and Nonce) to the file you are retrieving! So, the GET should be treated as *successful*, although you discard the copy of the file that you just got. -- Bill Cheng // bill.cheng@usc.edu On 4/20/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > From your response to the question below, what I understand from > > this particular test case is that there is only one copy (same nonce, > sha1, > > filename) of the file "blondie1.mp3" stored in the network when user > invoked > > > > in 1st window, type: > > store blondie1.mp3 1 title="Heart of Glass" > > > > At this point, node at 3rd window should have this file in its cache > > space. After it searches and gets 3 responses, no matter if user does > "get 1 > > 2 or 3", this should always succeed and hence make a copy of the file in > CWD > > because: > > 1. If the file corresponding to 'get' number is a local copy, it > > should simply move it to permanent space. > > 2. If the file comes from the other two nodes, it should check > > before attempting to store and find that it already has a copy of the > same > > file, and discard the response. > > So the permanent space should NEVER be exceeded in this case since > > there is only one file in the network, and a node MUST always have only > one > > copy of the file. I suspect that the person who got the storage exceeded > > must not check for file duplicate before he attempts to store the > received > > file. Please let me know what you think. > > Your analysis is correct! I'm reverting the grading guidelines > back to the previous version! > > Sorry about my confusion! > -- > Bill Cheng // bill.cheng@usc.edu > > > > > > > -----Original Message----- > From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] > Sent: Friday, April 20, 2007 2:40 PM > To: cs551@merlot.usc.edu > Subject: Re: CS551 - Final Project - Question on Grading guidelines > > Someone wrote: > > > For Test Case (A) Beacon-only networks -> (2) Network with 3 beacon > nodes > > > > The configuration parameters are set with CacheSize=600 and > PermSize=600 > > (These are in KB). > > > > For the following part of the test case: > > > > in 3rd window, type: > > (+1 points) > > search filename=blondie1.mp3 > > should get three responses, then do > > (+1 points) > > get 1 > > "blondie1.mp3" should be created in the > > current directory > > > > in the 4th window, type at the UNIX prompt: > > "ls -l blondie1.mp3" > > (file size should be 474736) > > "openssl sha1 blondie1.mp3" > > (should be > > > ffd3b197e1c0f0c27e7bdc219f553a3e6b139dfb) > > "rm blondie1.mp3" > > > > > > > > =========================> At this point 400 KB of the Perm space is > used > > up; which mean approx 200 KB of perm space is available. If the > following > > "get" commands are executed, then the node at window 3 will not be > able to > > save the file in the permanent area (due to lack of free space 400KB > > > > 200KB) and hence will not be able to provide the file to the user. So > are > > the subsequent steps correct? > > > > in 3rd window, type: > > (+1 points) > > get 2 > > "blondie1.mp3" should be created in the > > current directory > > > > in the 4th window, type at the UNIX prompt: > > "ls -l blondie1.mp3" > > "openssl sha1 blondie1.mp3" > > "rm blondie1.mp3" > > > > in 3rd window, type: > > (+1 points) > > get 3 > > "blondie1.mp3" should be created in the > > current directory > > > > in the 4th window, type at the UNIX prompt: > > "ls -l blondie1.mp3" > > "openssl sha1 blondie1.mp3" > > "rm blondie1.mp3" > > You are correct. I have just changed the grading guidelines > to say that "get 2" should not succeed and I removed "get 3" > and move the +1 point to "get 1". > > Thanks for catching the bug! > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 15:14:57 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3KMEvD5004557 for ; Fri, 20 Apr 2007 15:14:57 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3KMEni8017231 for ; Fri, 20 Apr 2007 15:14:49 -0700 Message-Id: <200704202214.l3KMEni8017231@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Question on Grading guidelines Date: Fri, 20 Apr 2007 15:14:49 -0700 From: william@bourbon.usc.edu Someone wrote: > From your response to the question below, what I understand from > this particular test case is that there is only one copy (same nonce, sha1, > filename) of the file "blondie1.mp3" stored in the network when user invoked > > in 1st window, type: > store blondie1.mp3 1 title="Heart of Glass" > > At this point, node at 3rd window should have this file in its cache > space. After it searches and gets 3 responses, no matter if user does "get 1 > 2 or 3", this should always succeed and hence make a copy of the file in CWD > because: > 1. If the file corresponding to 'get' number is a local copy, it > should simply move it to permanent space. > 2. If the file comes from the other two nodes, it should check > before attempting to store and find that it already has a copy of the same > file, and discard the response. > So the permanent space should NEVER be exceeded in this case since > there is only one file in the network, and a node MUST always have only one > copy of the file. I suspect that the person who got the storage exceeded > must not check for file duplicate before he attempts to store the received > file. Please let me know what you think. Your analysis is correct! I'm reverting the grading guidelines back to the previous version! Sorry about my confusion! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Friday, April 20, 2007 2:40 PM To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Question on Grading guidelines Someone wrote: > For Test Case (A) Beacon-only networks -> (2) Network with 3 beacon nodes > > The configuration parameters are set with CacheSize=600 and PermSize=600 > (These are in KB). > > For the following part of the test case: > > in 3rd window, type: > (+1 points) > search filename=blondie1.mp3 > should get three responses, then do > (+1 points) > get 1 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > (file size should be 474736) > "openssl sha1 blondie1.mp3" > (should be > ffd3b197e1c0f0c27e7bdc219f553a3e6b139dfb) > "rm blondie1.mp3" > > > > =========================> At this point 400 KB of the Perm space is used > up; which mean approx 200 KB of perm space is available. If the following > "get" commands are executed, then the node at window 3 will not be able to > save the file in the permanent area (due to lack of free space 400KB > > 200KB) and hence will not be able to provide the file to the user. So are > the subsequent steps correct? > > in 3rd window, type: > (+1 points) > get 2 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > "openssl sha1 blondie1.mp3" > "rm blondie1.mp3" > > in 3rd window, type: > (+1 points) > get 3 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > "openssl sha1 blondie1.mp3" > "rm blondie1.mp3" You are correct. I have just changed the grading guidelines to say that "get 2" should not succeed and I removed "get 3" and move the +1 point to "get 1". Thanks for catching the bug! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 14:39:45 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3KLdjba001585 for ; Fri, 20 Apr 2007 14:39:45 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3KLdbHa016224 for ; Fri, 20 Apr 2007 14:39:37 -0700 Message-Id: <200704202139.l3KLdbHa016224@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Question on Grading guidelines Date: Fri, 20 Apr 2007 14:39:37 -0700 From: william@bourbon.usc.edu Someone wrote: > For Test Case (A) Beacon-only networks -> (2) Network with 3 beacon nodes > > The configuration parameters are set with CacheSize=600 and PermSize=600 > (These are in KB). > > For the following part of the test case: > > in 3rd window, type: > (+1 points) > search filename=blondie1.mp3 > should get three responses, then do > (+1 points) > get 1 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > (file size should be 474736) > "openssl sha1 blondie1.mp3" > (should be > ffd3b197e1c0f0c27e7bdc219f553a3e6b139dfb) > "rm blondie1.mp3" > > > > =========================> At this point 400 KB of the Perm space is used > up; which mean approx 200 KB of perm space is available. If the following > "get" commands are executed, then the node at window 3 will not be able to > save the file in the permanent area (due to lack of free space 400KB > > 200KB) and hence will not be able to provide the file to the user. So are > the subsequent steps correct? > > in 3rd window, type: > (+1 points) > get 2 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > "openssl sha1 blondie1.mp3" > "rm blondie1.mp3" > > in 3rd window, type: > (+1 points) > get 3 > "blondie1.mp3" should be created in the > current directory > > in the 4th window, type at the UNIX prompt: > "ls -l blondie1.mp3" > "openssl sha1 blondie1.mp3" > "rm blondie1.mp3" You are correct. I have just changed the grading guidelines to say that "get 2" should not succeed and I removed "get 3" and move the +1 point to "get 1". Thanks for catching the bug! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 13:45:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3KKjwOt029085 for ; Fri, 20 Apr 2007 13:45:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3KKjo4p015128 for ; Fri, 20 Apr 2007 13:45:50 -0700 Message-Id: <200704202045.l3KKjo4p015128@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt on GET message Date: Fri, 20 Apr 2007 13:45:50 -0700 From: william@bourbon.usc.edu Someone wrote: > Does this mean that the node replies to its own SEARCH msg? It must! A node also replies to its own STATUS, GET, DELETE, etc. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Thursday, April 19, 2007 7:46 am Subject: Re: Doubt on GET message To: cs551@merlot.usc.edu > Someone wrote: > > > What if the node that originates the GET request is the node > that has > > that FileID? Should it bother flooding anyway? > > It's optional then. If you separate the flood from the > "special lookup", then you can do these in parallel. > -- > Bill Cheng // bill.cheng@usc.edu > > > > On Apr 13, 2007, at 3:03 PM, william@bourbon.usc.edu wrote: > > Someone wrote: > > > >> Given the uniqueness of FileID, if a node finds that he > >> has the file requested by GET request, is it OK for > >> this node to stop forwarding this GET request and > >> just send back the file through GET_R? In other words, > >> could we say that if there's a match, this will be the > >> only match in the network so we no longer need to > >> forward this GET (other nodes can't have this file)? > > > > To provide "plausible deniability", you should continue to > > flood. (This should also simply your code, slightly.) > > -- > > Bill Cheng // bill.cheng@usc.edu > > usc/> Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 20 13:45:05 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3KKj5Pk029070 for ; Fri, 20 Apr 2007 13:45:05 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3KKivQX015101 for ; Fri, 20 Apr 2007 13:44:57 -0700 Message-Id: <200704202044.l3KKivQX015101@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question about SEARCH responce Date: Fri, 20 Apr 2007 13:44:57 -0700 From: william@bourbon.usc.edu Someone wrote: > Is the FileID generated at the time of receiving a SEARCH msg? You can do it when you are about to generate a SEARCH response. You only need to do it for files that got a hit in a search. > If yes is it persistant data and should be saved on disk or is it > just valid between two searches and does not need to be written > to disk? If you do it when you generate a SEARCH response, then you don't need to make it persistent. Alternatively, you can generate FileIDs when you STORE a file. In that case, it has to be persistent. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Sunday, April 15, 2007 8:48 am Subject: Re: Question about SEARCH responce To: cs551@merlot.usc.edu > Someone wrote: > > > As specs shows > > > > [1] FileID=*02adefc1dfc97a082fa18a5ef1e8c487259b7fb4 > > *FileName=foo > > FileSize=123 > > SHA1=b83a758fecbefcd3ea547fbf0f9a97eba0ea984c > > Nonce=*01b7a1bd6f169dde22518a865ab2f44c70fcab82 > > *Keywords=key1 key2 key3 ..... > > > > these two file ID's are different, > > so is the FileID is the ID of the file, which is got by > calling GetUOID > > at the time of sending search responce? as it is different > than the > > Nonce fieald os metadata? > > Correct. Please see: > > http://merlot.usc.edu/cs551-s07/projects/final.html#nonce > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 22:37:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5b9Yk021831 for ; Thu, 19 Apr 2007 22:37:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5b3CF004851 for ; Thu, 19 Apr 2007 22:37:03 -0700 Message-Id: <200704200537.l3K5b3CF004851@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: certificate field Date: Thu, 19 Apr 2007 22:37:03 -0700 From: william@bourbon.usc.edu Someone wrote: > There is a certificate field in the store msg (also in Get > response). Should we transfer this field as binary or text > representation of the SSL generated certificate? The certificate is an ASCII file (as far as I know). So, you just put the file in the certificate field. I'm not sure what binary data you are referring to. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 22:34:38 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5Ycrk021660 for ; Thu, 19 Apr 2007 22:34:38 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5YXV2004789 for ; Thu, 19 Apr 2007 22:34:33 -0700 Message-Id: <200704200534.l3K5YXV2004789@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Press Ctrl - C to interrupt waiting Date: Thu, 19 Apr 2007 22:34:33 -0700 From: william@bourbon.usc.edu Someone wrote: > The A(1) section of the grading guideline says: > > in the 1st window, continue and type: > (+1 points) > search filename=foo > should get no response, type c to interrupt > (+1 points) > search sha1hash=abcdef > should get no response, type c to interrupt > (+1 points) > search keywords=hello > should get no response, type c to interrupt > (+1 points) > search keywords="chess.jpg" > should get no response, type c to interrupt > type: > (+1 points) > search filename=chess.jpg > should get a response, type: > (+1 points) > search sha1hash=f7917fe4976d2c24f225bc4b6c2334e554b91c28 > should get a response, type: > (+2 points) > search keywords=mp3 > > Is it OK for the program to immediately return to prompt since > it can detect that there's no neighbor? (neighbor list len == 0). Yes. > If yes, that the program in this case will *not* require a -C > to get back to user. Right, but a should not hurt either. When you are sitting at the command prompt, pressing should only cause your program to print the prompt again. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 22:16:06 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5G6vS020203 for ; Thu, 19 Apr 2007 22:16:06 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3K5G0oj004452 for ; Thu, 19 Apr 2007 22:16:01 -0700 Message-Id: <200704200516.l3K5G0oj004452@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Status Response Date: Thu, 19 Apr 2007 22:16:00 -0700 From: william@bourbon.usc.edu Someone wrote: > The status message is meant for 2 types of functions Neighbors > and Files. Its fairly straightforward to differenctiate the > functionalities of the message type but the Respose message type > header is a bit ambiguous in the sense that machine > detecting the status respose cannot make it out as to what > exactly is the type of status responce its accepting( FILES of > NEIGHBORS) > > We have used a low level bit to indicate the same. However the > specs have no mention about this . I wanted to know what > should be the exact method of differentiating the respose > messages of STATUS. Your node can only run one user command at a time. So, it should know what type of status response it's waiting for. Right? -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 22:02:30 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3K52Ti6019060 for ; Thu, 19 Apr 2007 22:02:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3K52O0i004113 for ; Thu, 19 Apr 2007 22:02:24 -0700 Message-Id: <200704200502.l3K52O0i004113@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Storage of FileSpec and SignedFileSpec Date: Thu, 19 Apr 2007 22:02:24 -0700 From: william@bourbon.usc.edu Someone wrote: > I am not hardcoding it. > I ll explain u in detail what i am doing. > > I have a counter which goes on increasing. > Whenever i create tempfiles i put them inside homeDir/files/ > but with the name > > 1)temp1.data,temp1.meta,temp1.cert(whenever i am gettin a get > or store message) > temp2.data... > > 2)temp5.meta,temp6.meta...(when ever i get a search response.. > which has many meta files) > > 3)"i would also like to have temp10.txt,temp11.txt "(for > filespec and signed filespec)- so the name wud not be hardcoded. > > So basically i am havin all files in the /homeDir/files/ > And all the temp*.* i would delete in the bbeginning itself. > > I have a different counter for permanent and cache files. > The permanent and cache data files would be as 1.data,3.data..... > > So i am taking care that i never use a filename that i use once. > Is this approach rite? As I've mentioned before, it's fine since you are cleaning up. (But I would do things differently for temporary files. I would call mkstemp() and create these files in /tmp. Of course you don't have to take my approach.) -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Thursday, April 19, 2007 12:02 pm Subject: Re: Reg Storage of FileSpec and SignedFileSpec To: cs551@merlot.usc.edu > Someone wrote: > > > Can we store the FileSpec and signed filespec any wer in the > homedir? > can we store it also in the /files/ directory? > > You should use /tmp and call mkstemp() or equivalent to get > a filename. Please do *not* hardcode the filename. Please > also remember to delete the file after you are done. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 12:01:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3JJ1wig003767 for ; Thu, 19 Apr 2007 12:01:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3JJ1s5Z030931 for ; Thu, 19 Apr 2007 12:01:54 -0700 Message-Id: <200704191901.l3JJ1s5Z030931@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Storage of FileSpec and SignedFileSpec Date: Thu, 19 Apr 2007 12:01:54 -0700 From: william@bourbon.usc.edu Someone wrote: > Can we store the FileSpec and signed filespec any wer in the homedir? > can we store it also in the /files/ directory? You should use /tmp and call mkstemp() or equivalent to get a filename. Please do *not* hardcode the filename. Please also remember to delete the file after you are done. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 10:03:48 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3JH3mJU026762 for ; Thu, 19 Apr 2007 10:03:48 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3JH3iTK029476 for ; Thu, 19 Apr 2007 10:03:44 -0700 Message-Id: <200704191703.l3JH3iTK029476@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Get Response Query Date: Thu, 19 Apr 2007 10:03:44 -0700 From: william@bourbon.usc.edu Someone wrote: > You said: > > In the mini-filesystem, you should always use the information > > in the original file metadata. Therefore, it should be > > "Blondie.mp3". > > Does this mean this is the name saved in the metadata file > while the file contents are saved as index.data where index > is an integer? Yes. If the metadata file is "#.meta" where # is an integer, the corresponding file should be "#.data". Once the metadata of a file is created, it just gets copied around and never modified. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Thursday, April 19, 2007 7:50 am Subject: Re: CS551 - Final Project - Get Response Query To: cs551@merlot.usc.edu > Someone wrote: > > > For the following the following scenario: > > > > There is a file named "Blondie.mp3" in the servant network. > > > > A user fires the user command "get 2 XYZ.mp3". Once the file > reaches the > > user's node (as a consequence of a successful GetResponse), it > is stored as > > XYZ.mp3 in the current working directory of the user. > > > > Now, the node also has to store the file in its permanent > area. In this > > case, will the FileName be "Blondie.mp3" or "XYZ.mp3" (for the > meta data)? > > I'm asking because this will affect future searches by filename. > > In the mini-filesystem, you should always use the information > in the original file metadata. Therefore, it should be > "Blondie.mp3". > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 07:49:13 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEnDiX016334 for ; Thu, 19 Apr 2007 07:49:13 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEnAfA027850 for ; Thu, 19 Apr 2007 07:49:10 -0700 Message-Id: <200704191449.l3JEnAfA027850@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS551 - Final Project - Get Response Query Date: Thu, 19 Apr 2007 07:49:10 -0700 From: william@bourbon.usc.edu Someone wrote: > For the following the following scenario: > > There is a file named "Blondie.mp3" in the servant network. > > A user fires the user command "get 2 XYZ.mp3". Once the file reaches the > user's node (as a consequence of a successful GetResponse), it is stored as > XYZ.mp3 in the current working directory of the user. > > Now, the node also has to store the file in its permanent area. In this > case, will the FileName be "Blondie.mp3" or "XYZ.mp3" (for the meta data)? > I'm asking because this will affect future searches by filename. In the mini-filesystem, you should always use the information in the original file metadata. Therefore, it should be "Blondie.mp3". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 07:46:05 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEk5fU016164 for ; Thu, 19 Apr 2007 07:46:05 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEk2cv027748 for ; Thu, 19 Apr 2007 07:46:02 -0700 Message-Id: <200704191446.l3JEk2cv027748@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt on GET message Date: Thu, 19 Apr 2007 07:46:02 -0700 From: william@bourbon.usc.edu Someone wrote: > What if the node that originates the GET request is the node that has > that FileID? Should it bother flooding anyway? It's optional then. If you separate the flood from the "special lookup", then you can do these in parallel. -- Bill Cheng // bill.cheng@usc.edu On Apr 13, 2007, at 3:03 PM, william@bourbon.usc.edu wrote: > Someone wrote: > >> Given the uniqueness of FileID, if a node finds that he >> has the file requested by GET request, is it OK for >> this node to stop forwarding this GET request and >> just send back the file through GET_R? In other words, >> could we say that if there's a match, this will be the >> only match in the network so we no longer need to >> forward this GET (other nodes can't have this file)? > > To provide "plausible deniability", you should continue to > flood. (This should also simply your code, slightly.) > -- > Bill Cheng // bill.cheng@usc.edu usc/> Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 19 07:35:26 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEZQrt015212 for ; Thu, 19 Apr 2007 07:35:26 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3JEZN9l027599 for ; Thu, 19 Apr 2007 07:35:23 -0700 Message-Id: <200704191435.l3JEZN9l027599@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Delete Date: Thu, 19 Apr 2007 07:35:23 -0700 From: william@bourbon.usc.edu Someone wrote: > When we are doin delete the "delete FileName=..... SHA1=....... > Nonce=......" > should it be in the same format rite? can we allow for the user > to give in differnt or in lowercase(filename or FILENAME).. > or exactly the format specified by the spec as above? Please only allow the precise format specified in the spec. Thanks! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 18 11:20:30 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3IIKUYc016187 for ; Wed, 18 Apr 2007 11:20:30 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3IIKTT2015995 for ; Wed, 18 Apr 2007 11:20:29 -0700 Message-Id: <200704181820.l3IIKTT2015995@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Storage. Date: Wed, 18 Apr 2007 11:20:29 -0700 From: william@bourbon.usc.edu Someone wrote: > Suppose "get 2 abcd.jpg" is the command that is given and the > actual filename is "blondie1.mp3" and it is stored in my > permanent as 3.data... So we shud create the file abcd.jpg in the > current working directory rite? Yes. > or shud we change the name of 3.data to abcd.jpg? No. Please remember that the mini-filesystem belongs to the node and the current working directory belongs to the user running the node. > While it is in cache...we will move it to permanent Yes. > - supose in cache it is 2.data shud i move it as 2.data > or abcd.jpg to permanent... Well, moving from "2.data" to "2.data" is a no-op. You just need to change a bit somewhere in your mini-filesystem to indicate that the file is in permanent space now. > my guess is only in current working > directory we will change name and in permanent and cache it > always remains as (fileno).data. i just wanted to ask if it was > correct? I don't really think of it as "changing name". It's saving the file as a specified filename. Just like when you download a file from the internet, the browser often asks you to give the file a name as soon as the file leaves the browser space and into your local directory. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 18 07:50:24 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEoOkw031384 for ; Wed, 18 Apr 2007 07:50:24 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEoOP1013540 for ; Wed, 18 Apr 2007 07:50:24 -0700 Message-Id: <200704181450.l3IEoOP1013540@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Request for exchange of submission dates Date: Wed, 18 Apr 2007 07:50:24 -0700 From: william@bourbon.usc.edu Someone wrote: > I was just curious to know and would like to request, if we can > the submission dates of Homework#3 and Final Project#2 swapped, > that would help us in ensuring that we completely finish Final > Project#2. I'm sorry, but as I've mentioned before, I will only extend the project deadline if there is a major error in the spec or extended outage on nunki within 48 hours of the submission deadline. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 18 07:46:27 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEkR1m031056 for ; Wed, 18 Apr 2007 07:46:27 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEkS0U013476 for ; Wed, 18 Apr 2007 07:46:28 -0700 Message-Id: <200704181446.l3IEkS0U013476@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question about buffer limits Date: Wed, 18 Apr 2007 07:46:28 -0700 From: william@bourbon.usc.edu Someone wrote: > Is it reasonable to assume that the only messages that will ever be > larger than the 8192 byte buffer limit are STORE and GET RESPONSE, > and that within those messages only the actual file data will ever be > larger than 8192 bytes (i.e. not the metadata or certificate)? Yes. I think metadata and certificates are much smaller than that. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 18 07:42:47 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEglup030726 for ; Wed, 18 Apr 2007 07:42:47 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3IEgl5x013420 for ; Wed, 18 Apr 2007 07:42:47 -0700 Message-Id: <200704181442.l3IEgl5x013420@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Bitvectors and keywords Date: Wed, 18 Apr 2007 07:42:47 -0700 From: william@bourbon.usc.edu Someone wrote: > If i keep a multimap structure for storing the bitvectors created > by the keywords in the metadata file, so as every bitvectors > maps a set of keywords and the file containing them then i really > dont feel any need of keeping the Keywords index > structures to quiken the search operation coz then we would be > incorporating redundancy in the logic and code. > (END to END argument... ;) :) Please remember that one of the main "escape clause" for the end-to-end argument is efficiency. And the main point of having an index structure is to introduce redundancy to speed up certain operations. >From your description, even though it's an improvement over traditional data structures, it doesn't scale as well as the bit-vector approach. > Do i still require extra keywords index then?? Yes. Please see the grading guidelines on "minu points". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 20:08:24 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3I38O2I008284 for ; Tue, 17 Apr 2007 20:08:24 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3I38Q4L005714 for ; Tue, 17 Apr 2007 20:08:26 -0700 Message-Id: <200704180308.l3I38Q4L005714@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Questions on test guideline Date: Tue, 17 Apr 2007 20:08:26 -0700 From: william@bourbon.usc.edu Someone wrote: > I have more questions on test guideline. > > 1. In (B)(1)(c), it says, > > search filename=f1.wav > should get several responses, find a copy of the > file that is cached, let's call this x, then type: > > (+1 points) > get [x] > > How can we know which search response corresponds to a 'cached' copy? > I am also not sure why we only need to get a file from a cached copy. You are correct. I've changed it in a similar fashion as before. > 2. In the 'Minus points' section, it says, > > if did not use O(n) algorithm for deserialization : -3 points > > What does this mean? This was an old requirement! It is deleted now. Thanks for catching these! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 19:43:03 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2h36l006176 for ; Tue, 17 Apr 2007 19:43:03 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2h5tT005331 for ; Tue, 17 Apr 2007 19:43:05 -0700 Message-Id: <200704180243.l3I2h5tT005331@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: store msg recap Date: Tue, 17 Apr 2007 19:43:05 -0700 From: william@bourbon.usc.edu Someone wrote: > I know you have asnwered this before but I just wanted > confirmation (yes/no) > > When a user enters a store command, if the file cant be copied to > the permanent storage then the store msg isnt flooded even if > TTL>0. Correct. This is stated in item (6) of Summary of Basic Message Types section. > A node getting a store msg forwards the store msg > probabilistically (by flipping a coin) for each neighbor (if TT > still > 0) even if it cant store the msg in its own cache. Correct. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 19:36:00 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2a01H005545 for ; Tue, 17 Apr 2007 19:36:00 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2a2N6005249 for ; Tue, 17 Apr 2007 19:36:02 -0700 Message-Id: <200704180236.l3I2a2N6005249@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About delete Date: Tue, 17 Apr 2007 19:36:02 -0700 From: william@bourbon.usc.edu Someone wrote: > I am not clear how the delete works? > When a node gets a delete packet it verifies it with all the .pem files > he has in his home directory and then matches the contents(file name, > sha1, nonce) provided in the delete packet > nad then deletes the file.. > is it the right flow isn't it very inefficient or am I missing something? This is inefficient. If you look at the sample signed File Spec: http://merlot.usc.edu/cs551-s07/projects/cert/signed.txt You should see that the following 3 lines are visible to your program: FileName=... SHA1=... Nonce=... So, you can extract the FileName, SHA1, and Nonce values from a signed File Spec *before* to try to verify any digital signatures. Then you can use, say, the SHA1 index structure to find all files with the same SHA1 files, open their metadata files and find a subset that has the same Noce (hopefullly, there is only one such file since you should not have duplicates), then apply the corresponding certificate file to verify digital signature. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 19:28:23 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2SNV7005056 for ; Tue, 17 Apr 2007 19:28:23 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3I2SP44005176 for ; Tue, 17 Apr 2007 19:28:25 -0700 Message-Id: <200704180228.l3I2SP44005176@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Storage. Date: Tue, 17 Apr 2007 19:28:25 -0700 From: william@bourbon.usc.edu Someone wrote: > And we have to use to set the name for the file in > CWD only rite? Yes. > or shud the filename be set for the file that is > (either existing or moved from cache) in Permanent? I'm not sure what you meant. Please remember that cache or permanent space is under HomeDir. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Tuesday, April 17, 2007 4:08 pm Subject: Re: Reg Storage. To: cs551@merlot.usc.edu > Someone wrote: > > > If my understanding was right, we shud not have any file that is > > duplicated(in the sense the metadata is same..incl.nonce) in the > > permanent or cache.(That is the file shud be in either of the > > storage and not in both of them). > > Correct. > > > Then in that case what shud we do > > 1)if the node has the file in its cache and u search for the same > > file and get it from its own cache. > > In the spec it is given that it shud be moved to Permanent. But > > shud it also be copied into CWD? > > It should. The reason a user enters a GET command is to get > a file into the current working directory. ("Moving" the > file between cache and permanent spaces is just a separate > and independent thing your node does.) > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 16:08:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HN8ICe021677 for ; Tue, 17 Apr 2007 16:08:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HN8KFp003425 for ; Tue, 17 Apr 2007 16:08:20 -0700 Message-Id: <200704172308.l3HN8KFp003425@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Storage. Date: Tue, 17 Apr 2007 16:08:20 -0700 From: william@bourbon.usc.edu Someone wrote: > If my understanding was right, we shud not have any file that is > duplicated(in the sense the metadata is same..incl.nonce) in the > permanent or cache.(That is the file shud be in either of the > storage and not in both of them). Correct. > Then in that case what shud we do > 1)if the node has the file in its cache and u search for the same > file and get it from its own cache. > In the spec it is given that it shud be moved to Permanent. But > shud it also be copied into CWD? It should. The reason a user enters a GET command is to get a file into the current working directory. ("Moving" the file between cache and permanent spaces is just a separate and independent thing your node does.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 15:36:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMa8Rb019174 for ; Tue, 17 Apr 2007 15:36:08 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMaBRf003021 for ; Tue, 17 Apr 2007 15:36:11 -0700 Message-Id: <200704172236.l3HMaBRf003021@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: bit vector on keywords Date: Tue, 17 Apr 2007 15:36:11 -0700 From: william@bourbon.usc.edu Someone wrote: > If the user enters: > > store chess.jpg 1 categories="audio mp3" artist="Blondie" > > Should we compute bit-vector for, in this case, each of the > 6 keywords above *ONE BY ONE*, or for the following > concatenated, quote-free string in *ONE PASS* ? > > "category audio mp3 artisti Blondie" One by one and then OR'ed the results together. Please remember to make everything lowercase before you do compute the bit-vector for a keyword. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 15:34:07 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMY7uA019006 for ; Tue, 17 Apr 2007 15:34:07 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMYA4X002979 for ; Tue, 17 Apr 2007 15:34:10 -0700 Message-Id: <200704172234.l3HMYA4X002979@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Test case 1 Date: Tue, 17 Apr 2007 15:34:10 -0700 From: william@bourbon.usc.edu Someone wrote: > From your response to the question below, when a node has no file, > if we do not create an entry for that node in the .out file, how can the > grader know if the status command works properly since there is no evidence > of node actually responded. I thought the spec said if a node has no file, > we should at least print "hostname:port has no file". Sorry! I completely forgot about that. You are correct. You should follow the spec! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Tuesday, April 17, 2007 11:07 AM To: cs551@merlot.usc.edu Subject: Re: Test case 1 Someone wrote: > I have some doubt about test case 1. In this case, > there's no files in the network, should we create > an empty .out file? > > PS. Test case 1: > > (1) Network with 1 beacon node : +28 points > autoshutdown=600 sec, keepalive timeout=10 sec, > msglifetime=15 sec > CacheProb=0, StoreProb=1, NeighborStoreProb=1 > CacheSize=10, PermSize=10 > > (a) in one window, do: > ~csci551b/bin/resetf2 > /bin/rm b3-n??.ini ??.out > cp ~/grading/guidelines/final2/b3-n??.ini . > cp ~csci551b/public/final2/*.gif . > cp ~csci551b/public/final2/*.jpg . > cp ~csci551b/public/final2/*.mp3 . > ./sv_node b3-n00.ini > (+2 points) > (wait 5 seconds, at the commandline prompt, type > "status files 1 00.out", look at "00.out") Correct. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 15:29:45 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMTj4C018530 for ; Tue, 17 Apr 2007 15:29:45 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HMTmGe002901 for ; Tue, 17 Apr 2007 15:29:48 -0700 Message-Id: <200704172229.l3HMTmGe002901@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on test guideline Date: Tue, 17 Apr 2007 15:29:48 -0700 From: william@bourbon.usc.edu Someone wrote: > I have some question in the testing guideline. > > 1. There are several places where it says to wait for nodes to > autoshutdown and then followed by "status" command. Is this correct? > > *wait* for all nodes to autoshutdown > > (+2 points) > in 1st window (node *00), type: > "status neighbors 5 00.out", look at "00.out", and > make sure the network looks corrct. > > *restart the network in the same order in windows 1 through 6* > > These are seen at: (B)(1)(a)-(c) You are correct. I've just swapped the "*wait* for all nodes to autoshutdown" lines with the "status neighbors" lines. > 2. One case has 'get [x]' where x refers to a number in the search > result matched to a fileID generated by a certain node. How can we know > which node generates which fileID, since this information is not displayed > anywhere? > > (+1 points) > > search filename=earc.jpg > should get 5 responses, > look for the FileID that matches the FileID at > node *02, let's call this x, then type: > (+1 points) > get [x] > > These are seen twice at: (B)(1)(a) You are absolutely, there is no way to find out which FileID corresponds to which node for these in (B.1.a). > From the context of this test case, it does not seem > to matter which node we get the file from as long as it is > the same copy. Then can we just get any one of the file from > the search response? I have changed them to say "get 1", "get 2", etc. Thanks for catching these bugs! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 11:07:14 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HI7EUE030132 for ; Tue, 17 Apr 2007 11:07:14 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HI7HoY032348 for ; Tue, 17 Apr 2007 11:07:17 -0700 Message-Id: <200704171807.l3HI7HoY032348@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Test case 1 Date: Tue, 17 Apr 2007 11:07:17 -0700 From: william@bourbon.usc.edu Someone wrote: > I have some doubt about test case 1. In this case, > there's no files in the network, should we create > an empty .out file? > > PS. Test case 1: > > (1) Network with 1 beacon node : +28 points > autoshutdown=600 sec, keepalive timeout=10 sec, > msglifetime=15 sec > CacheProb=0, StoreProb=1, NeighborStoreProb=1 > CacheSize=10, PermSize=10 > > (a) in one window, do: > ~csci551b/bin/resetf2 > /bin/rm b3-n??.ini ??.out > cp ~/grading/guidelines/final2/b3-n??.ini . > cp ~csci551b/public/final2/*.gif . > cp ~csci551b/public/final2/*.jpg . > cp ~csci551b/public/final2/*.mp3 . > ./sv_node b3-n00.ini > (+2 points) > (wait 5 seconds, at the commandline prompt, type > "status files 1 00.out", look at "00.out") Correct. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 11:00:53 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HI0r6X029479 for ; Tue, 17 Apr 2007 11:00:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HI0u6g032200 for ; Tue, 17 Apr 2007 11:00:56 -0700 Message-Id: <200704171800.l3HI0u6g032200@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: problem in fgets Date: Tue, 17 Apr 2007 11:00:56 -0700 From: william@bourbon.usc.edu Hi, I'm just forwarding an e-mail from a student. This may explain the previously mentioned bug. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Mon, 16 Apr 2007 23:51:16 -0700 To: william@bourbon.usc.edu Subject: RE: problem in fgets Hi Professor, I was facing a similar problem and found out that the reason was that at one part of the code, I was closing index 0. and that is stdin. :-) Just thought to share with you. Regards, > -----Original Message----- > From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] > Sent: Thursday, March 22, 2007 9:55 PM > To: cs551@merlot.usc.edu > Subject: Re: problem in fgets > > Someone wrote: > > > we are trying to implement the command prompt using fgets by > > calling it repeatedly from a seperate thread. But it looks like > > that it doesnt wait for the user input and infinitely prints the > > command prompt. Kindly suggest the solution to this problem. Is > > this a problem related with Pthread implementation.. > > Hmm... This sounds familiar. I think I've heard it from > previous semesters. But they were all fixed; although I > don't recalling anyone telling me what the fix was. > > If fgets() doesn't work but fread() one byte at a time works, > you can write your own fgets() using fread() one byte at a time. > -- > Bill Cheng // bill.cheng@usc.edu > > > Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 10:58:31 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HHwV2N029298 for ; Tue, 17 Apr 2007 10:58:31 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HHwYQC032136 for ; Tue, 17 Apr 2007 10:58:34 -0700 Message-Id: <200704171758.l3HHwYQC032136@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj Part 2 Date: Tue, 17 Apr 2007 10:58:34 -0700 From: william@bourbon.usc.edu Someone wrote: > TO use the function called min(a,b) what header should we > include?? You have used that in the GetUOID function given in the > spec. The following should work: #define min(A,B) (((A)>(B)) ? (B) : (A)) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 10:52:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HHqwv3028818 for ; Tue, 17 Apr 2007 10:52:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HHr15r032016 for ; Tue, 17 Apr 2007 10:53:01 -0700 Message-Id: <200704171753.l3HHr15r032016@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: problem with deleting Date: Tue, 17 Apr 2007 10:53:01 -0700 From: william@bourbon.usc.edu Someone wrote: > In the process of deleting at the receiver node, while using popen, I am > getting the following error message at the side of another node while > deleting, could you please explain what is wrong here? > > *> Error reading S/MIME message > Error reading S/MIME message > > *> 3533:error:2107A088:PKCS7 routines:SMIME_read_PKCS7:no multipart body > failure:pk7_mime.c:255: > 3533:error:2107A088:PKCS7 routines:SMIME_read_PKCS7:no multipart body > failure:pk7_mime.c:255: OpenSSL often gives cryptic message when things are not going correctly. My guess is that you are not feeding "openssl" with the proper data. You should run the command by hand and make sure what you are running is a valid command. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 17 09:57:06 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3HGv6TO024341 for ; Tue, 17 Apr 2007 09:57:06 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3HGv9Pe031342 for ; Tue, 17 Apr 2007 09:57:09 -0700 Message-Id: <200704171657.l3HGv9Pe031342@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: proj 2 Date: Tue, 17 Apr 2007 09:57:09 -0700 From: william@bourbon.usc.edu Someone wrote: > Question 1: > I have a doubt in search response. I can not understand what > is FileID in file response. > > [1] FileID=02adefc1dfc97a082fa18a5ef1e8c487259b7fb4 > FileName=foo > FileSize=123 > SHA1=b83a758fecbefcd3ea547fbf0f9a97eba0ea984c > Nonce=01b7a1bd6f169dde22518a865ab2f44c70fcab82 > Keywords=key1 key2 key3 Please see the following for an explanation of FileID: http://merlot.usc.edu/cs551-s07/projects/final.html#nonce > Question2: > > I accidentally find out that i am getting correct hash of any > text file but getting wrong hash of .mp3 and .jpg file. Can > you tell me why this is happening ?. If you are getting correct hash values for ASCII text files but incorrect ones for binary files, then you probably are reading the input file using some functions that assumes the input file is an ASCII text file. One way you can check if this is happening is keep track of how many bytes of data you've read. When you are done with reading the file, check the total number of bytes you have read and that must be the same as the file size. If it's not, you know you are doing something wrong. There is a hexdump program at: ~csci551b/bin/hexdump You can use it to do a hexdump of the file you are having problem with and check it against what your code have read. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 16 16:06:53 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3GN6rfY004777 for ; Mon, 16 Apr 2007 16:06:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3GN6x3A022310 for ; Mon, 16 Apr 2007 16:06:59 -0700 Message-Id: <200704162306.l3GN6x3A022310@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: duplicate files detection and msg cache Date: Mon, 16 Apr 2007 16:06:59 -0700 From: william@bourbon.usc.edu Someone wrote: > As I understand, a node must no store duplicate file > (a file he already has) when receiving a store message. > Nonce can distinguish difference instance of the same > file, but how can we distinguish whether a file in store > message is what we already have, given that same > filename, keywords, sha1 are not enough to distinguish > two files? If you have a file in your mini-filesystem with the same filename, sha1, and nonce, as in your message, you should not store a copy of the file in your message. (I'm not sure exactly what you are asking.) > Secondly, if an intermediate node successfully forwards > message A (as part of flooding originated by this node's neighbor), > and later response to A back to the previous hop, should we > erase msg cache entry for A (since A's path is no longer > useful for routing)? No, it's not necessary. Since you already have a timeout mechanism to deal with this. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 16:22:38 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FNMcO9022443 for ; Sun, 15 Apr 2007 16:22:38 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FNMl0R009423 for ; Sun, 15 Apr 2007 16:22:47 -0700 Message-Id: <200704152322.l3FNMl0R009423@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt about Grading Guideline Date: Sun, 15 Apr 2007 16:22:47 -0700 From: william@bourbon.usc.edu Someone wrote: > I was talking about > (A) Beacon-only networks : +52 points (Note: *CacheProb is useless for > beacon-only networks*) > --(1) Network with 1 beacon node : +28 points On the node that issued the request, you must search its datastore. > One more thing I want to ask, why have you mentioned *CacheProb is > useless for beacon-only networks, *because I can not find any thing > which tells CacheProb is not for beacons. > Is it right? or IU have missed something...? Becuase it's fully connected. So, when you reply, the message goes directly to the requester. (I guess if you have some really strange timing in your routing implementation, may be this may not happen. But for the (A) cases, we set CacheProb to 0, then it wouldn't matter.) -- Bill Cheng // bill.cheng@usc.edu > william@bourbon.usc.edu wrote: > > Someone wrote: > > > > > As per grading guidelines the node should get search response when there > > > is only one node is up which is the search requester > > > that means the requester should also search for the required search > > > query in it's lists and reply back to the search request > > > is it right? > > > or I am missing something... ? > > > > I don't know what you are asking. Could you tell me which > > test (?.?.?) in the grading guidelines you are asking about? > > Please be as specific as possible and copy and paste what's > > in the grading guideslines should also help. > > -- > > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 16:10:00 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FNA0Bt021350 for ; Sun, 15 Apr 2007 16:10:00 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FNAAJ6009237 for ; Sun, 15 Apr 2007 16:10:10 -0700 Message-Id: <200704152310.l3FNAAJ6009237@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Part2 Date: Sun, 15 Apr 2007 16:10:10 -0700 From: william@bourbon.usc.edu Someone wrote: > Q1.When a node gets a "GET RESPONSE" which is not destined to > itself, and it is supposed to route it to somebody else,it has to > store the file it receives so that it can forward it to that > node.My question is ,is it OK for us to store the file in the > mini file system directory and then after forwarding the file > ,if it decides not to cache the fiile (CACHE PROB) we should > remove the fiile.So any other file cached at this node (due to a > later STORE REQ OR GET RESPONSE)in the mini file system will not > have an index which is not continuous with the previously stored > file.So is it Ok for us to do something like this?Or should the > index values in the mini file system be continuous? (i.e..,is it > ok for us to store files like 1.data 2.data 4.data) It depends on what you meant by "stroe the file in the mini-filesystem". If you are going to save the *temporary* file as, say, "3.data", and then delete it if you decide that you don't need it, then I don't think it's a good idea. The reason is that if something happens (like your node crashed), next time you start the node, there would be no way to tell if "3.data" is a good file or a temporary file, then you may be in trouble. But you can easily just call it "3tmp.data"! Then it's clear if it's temporary or not. If you decide to keep it, you just need to copy or rename it to "3.data". But the again, there is no requirement how you must do it. So, you should do what makes sense to you. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 09:17:29 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FGHTvQ021303 for ; Sun, 15 Apr 2007 09:17:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FGHdkr005973 for ; Sun, 15 Apr 2007 09:17:39 -0700 Message-Id: <200704151617.l3FGHdkr005973@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: [e2e] Simulator for wireless network Date: Sun, 15 Apr 2007 09:17:39 -0700 From: william@bourbon.usc.edu Hi, I'm just forwarding a post by S. Keshav to the e2e mailing list. He explained nicely why simulation of networks is difficult. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Sun, 15 Apr 2007 10:18:13 -0400 From: "S. Keshav" To: end2end-interest@postel.org Subject: Re: [e2e] Simulator for wireless network Durga, The discussions on this topic should convince you, I hope, that before using simulations, their role has to be clearly understood. For physical systems, such as planes, cars, and sailboats, the primary operational parameters are the laws of physics, which can be modeled to as great a degree of accuracy as desired. In other words, I can overlay a 2D or 3D grid on the underlying system and apply the laws of physics at each grid point. This is why it's possible to accurately design physical systems from computer simulations. For computer systems, where a single line of code can completely change the behavior of the system, one has to confront the fact that no simulation is ever going to be accurate. As has been pointed out already, the issue is not just of radio propagation modeling, which is hard enough, but the problem is that there are many layers of software that intervene from the receipt of a radio signal and a user- perceptible effect. A slight change in any of these can materially affect the result. For instance, a slight change in the card firmware can change the way in which packets are handed to the driver, which can change the timing at which packets are received by an application, which may result in user-perceptible audio effects for VOIP over WLAN. It is practically impossible to model these with any accuracy, and even if you do, a patch to the firmware, driver, OS, or application will invalidate your results. This is the reason why 'proof by simulation', for computer systems, at least, is farcical. Not only are simulators known to be buggy, but they are also simulating a system that is too loosely coupled to be adequately modeled. So, does this mean that simulation is useless? Not really. Simulations are useful in helping to form intuitions about the underlying system. They can also help explore the parameter space systematically. But, one has to realize that it is a coarse tool, and necessarily so. As long as you go in with your eyes open, simulations are a reasonable first step (but only the first step). keshav Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 09:03:10 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FG3AlG020205 for ; Sun, 15 Apr 2007 09:03:10 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FG3Lk0005818 for ; Sun, 15 Apr 2007 09:03:21 -0700 Message-Id: <200704151603.l3FG3Lk0005818@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt about Grading Guideline Date: Sun, 15 Apr 2007 09:03:21 -0700 From: william@bourbon.usc.edu Someone wrote: > As per grading guidelines the node should get search response when there > is only one node is up which is the search requester > that means the requester should also search for the required search > query in it's lists and reply back to the search request > is it right? > or I am missing something... ? I don't know what you are asking. Could you tell me which test (?.?.?) in the grading guidelines you are asking about? Please be as specific as possible and copy and paste what's in the grading guideslines should also help. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 09:00:41 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FG0fnL020035 for ; Sun, 15 Apr 2007 09:00:41 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FG0qVH005778 for ; Sun, 15 Apr 2007 09:00:52 -0700 Message-Id: <200704151600.l3FG0qVH005778@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: relay store message Date: Sun, 15 Apr 2007 09:00:52 -0700 From: william@bourbon.usc.edu Someone wrote: > Should relay/forward of store message also > be probabilistic? Sure! That's what NeighborStoreProb is about. Normally, we do deterministic flooding. But for STORE request messages, we need to do probabilistic flooding. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 08:53:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FFrInj019380 for ; Sun, 15 Apr 2007 08:53:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FFrSrh005662 for ; Sun, 15 Apr 2007 08:53:28 -0700 Message-Id: <200704151553.l3FFrSrh005662@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: FileID Date: Sun, 15 Apr 2007 08:53:28 -0700 From: william@bourbon.usc.edu Someone wrote: > Other than creating a search reply message, > at what other occasion will a node create > File ID for a file? None, if the above is your approach. You can also take another approach. When you save a file into cache or permanent space, you create the FileID. But now you have to save the FileID in the mini-filesystem. With the first approach, you can only keep the FileIDs in memory. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 15 08:48:04 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3FFm4SI019042 for ; Sun, 15 Apr 2007 08:48:04 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3FFmEml005567 for ; Sun, 15 Apr 2007 08:48:14 -0700 Message-Id: <200704151548.l3FFmEml005567@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question about SEARCH responce Date: Sun, 15 Apr 2007 08:48:14 -0700 From: william@bourbon.usc.edu Someone wrote: > As specs shows > > [1] FileID=*02adefc1dfc97a082fa18a5ef1e8c487259b7fb4 > *FileName=foo > FileSize=123 > SHA1=b83a758fecbefcd3ea547fbf0f9a97eba0ea984c > Nonce=*01b7a1bd6f169dde22518a865ab2f44c70fcab82 > *Keywords=key1 key2 key3 ..... > > these two file ID's are different, > so is the FileID is the ID of the file, which is got by calling GetUOID > at the time of sending search responce? as it is different than the > Nonce fieald os metadata? Correct. Please see: http://merlot.usc.edu/cs551-s07/projects/final.html#nonce -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 20:17:11 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3F3HBep030892 for ; Sat, 14 Apr 2007 20:17:11 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3F3HNdp006198 for ; Sat, 14 Apr 2007 20:17:23 -0700 Message-Id: <200704150317.l3F3HNdp006198@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Project 2 Date: Sat, 14 Apr 2007 20:17:23 -0700 From: william@bourbon.usc.edu Someone wrote: > Q1.If you hit a store request at a node and if the permanent size > is not sufficient to store the file,then should we still forward > the store request to the neighbors? If this is not the node that originates the STORE request, you should forward the STORE request. > Q2.Once it is decided that the file is to be cached at a node > when a store request comes or when the node is acting as an > intermediate node and it gets a GET RESPONSE,then if the CACHE > SIZE is not big enough to hold the file,then what should be done? You should not cache the file if the size of the file is larger than CacheSize. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 16:14:00 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3ENE0iY011730 for ; Sat, 14 Apr 2007 16:14:00 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3ENEDwH004334 for ; Sat, 14 Apr 2007 16:14:13 -0700 Message-Id: <200704142314.l3ENEDwH004334@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt on Search Date: Sat, 14 Apr 2007 16:14:13 -0700 From: william@bourbon.usc.edu Someone wrote: > I am a little confused with how a search is considered hit, so I > just want to clarify this. > 1. For a search to be considered 'hit', ALL search keywords must > match at least some but not necessarily all file keywords. For example, if > searching for "mp3 categories" and the file has "categories blondie mp3", > this is a hit. Correct. > I terms of operation, both bit-vectors should AND to 1, then > we can start comparing individual keyword. Actually, if you form the serarch bit-vector, let's call this S. And you AND it with a file bit-vector, let's call this F, if the outcome (S AND F) is exactly equal to S, then you need to compare individual keyword. > 2. If search keyword has at least a single keyword that is not in > the file keywords, it is not a hit. For example, "mp3 categories picture" is > not a hit of "categories blondie mp3". Correct. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 15:52:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMq9OA010084 for ; Sat, 14 Apr 2007 15:52:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMqNEZ004047 for ; Sat, 14 Apr 2007 15:52:23 -0700 Message-Id: <200704142252.l3EMqNEZ004047@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Questions on test guideline Date: Sat, 14 Apr 2007 15:52:23 -0700 From: william@bourbon.usc.edu Someone wrote: > 1. When I run '~csci551b/public/final2/scripts/setup-r789' it > returns 'Usage: setup-r789 $HOME'. So I believe this command needs an > argument. Can you please clarify what this command does? You should actually run: ~csci551b/public/final2/scripts/setup-r789 $HOME Your shell will replace $HOME by your actual home directory. > 2. For (B)(1)(a), I can't seem to find the file "earc.jpg" among > the test files. Is it related to not having command in (1) run? Hmm... It's at: ~csci551b/public/final2/earc.jpg Apparently, it was not included in ~csci551b/public/final2.tar. I've just updated it. Thanks for catching it. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 15:43:42 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMhf31009295 for ; Sat, 14 Apr 2007 15:43:41 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMht4N003952 for ; Sat, 14 Apr 2007 15:43:55 -0700 Message-Id: <200704142243.l3EMht4N003952@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Merlot Server down Date: Sat, 14 Apr 2007 15:43:55 -0700 From: william@bourbon.usc.edu Someone wrote: > I am unable to open the CSCI551-s07 site since 12:00 this afternoon. The server has not gone down. However,I have noticed that sometimes I cannot connect to it on my Mac (but have no problem with Windows XP). I noticed that if I just do a "ping merlot.usc.edu" on my Mac, then I was able to connect to the web server! Are you using a Mac? -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 15:40:57 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMevmC009074 for ; Sat, 14 Apr 2007 15:40:57 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EMfBSo003893 for ; Sat, 14 Apr 2007 15:41:11 -0700 Message-Id: <200704142241.l3EMfBSo003893@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Testing guideline Part(A) Date: Sat, 14 Apr 2007 15:41:11 -0700 From: william@bourbon.usc.edu Someone wrote: > I want to point out a few things in testing guideline. > > (A)(1)(a) > (1) search sha1hash=abcdef > - Is it OK if we do validity check on the size of hash, and > if it's not 20 bytes long, the program immediately returns without trying to > flood 'search'? Yes. > (2) search keywords=mp3 > should get a response > > - The stored keywords were "categories audio mp3 artist > Blondie". Here the search keyword (mp3) doesn't match ALL the file keywords > so we should not get a response, correct? No. You should get a hit. If your search command has N keywords, you must return all files that contains all these N keywords. (And not the other way around!) > (3) search keywords="audio mp3 blondie" > should get a response, then do > > - This is similar to (2). You should also get a hit for the file above. (You should also turn all the keywords to lowercase before doing the bit-vector and strcmp().) > (4) get 1 > > - This follows from (3). If (3) doesn't return anything, we > won't be able to 'get'. We could modify (3) such that it contains all > keywords so we can do 'get' here. Following from what I've mentioned, you should GET the file. > (A)(2)(a) > (1) store blondie1.mp3 1 title="Heart of Glass" > > in 2nd window, type: > "status files 1 01.out", look at "01.out", > every node should have this file, the metadata > should look like the following: > > - I just want to confirm that 01.out should contain 3 > duplicate metadata, one returned from each of the three node, correct? I think so. If I'm wrong, please let me know! > (2) search filename=blondie1.mp3 > should get a response, then do > > - I think we should get THREE responses here since the > previous 'store' causes each of the three nodes to have one copy (StoreProb > = 1). You are correct. > (3) store usctommy.gif 1 category="picture gif" > type "status files 1 00.out", look at "00.out", > every node should have this file, the metadata > should look like the following: > > - Just want to confirm that 00.out should now have 6 entries > of metadata, 3 for 'blondie.mp3' and 3 for 'usctommy.gif'. I think so. > (A)(2)(c) > (1) type "status files 1 00.out", look at "00.out", > (+2 points) > only node *00 should have all 7 files, > node *01 and node *02 should have about 50% of > these files since StoreProb=0.7 and > NeighborStoreProb=0.7 > > - With 7 files probabilistically flooded and stored, we > expect node *01 and node *02 to have 50% (3-4 files). However, each of these > file have the size of 1874 bytes but their CacheSize is only 5kB (according > to their .ini). So it will be able to cache at most 2 files at a time since > LRU algo. should be triggered when it attempts to cache the third file. I > think we should either increase the CacheSize or update the guideline to say > that only 2 files are expected so LRU alog. is consequently tested. You are correct. I've updated the grading guidelines. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 15:08:29 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EM8TLr006460 for ; Sat, 14 Apr 2007 15:08:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EM8gS0003497 for ; Sat, 14 Apr 2007 15:08:42 -0700 Message-Id: <200704142208.l3EM8gS0003497@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Index Date: Sat, 14 Apr 2007 15:08:42 -0700 From: william@bourbon.usc.edu Someone wrote: > What exactly should be the content of the index files?? > For example the file name_index should it contain the names > of all the files stored on that node?? Let's use the name_index as an example, so seaches will all based on the value of FileName in the file metadata. Think about the case where your node stores a lot of files. When your node comes up, how would you build the in-memory name index so you can satisfy search requests efficiently? If you scan through your mini-filesystem and open up each .meta file, you can build this in-memory name index. But clearly that's too slow. So, how can you speed it up? What you need is a disk-image of the in-memory name index. So, when your node comes up, you simply read the name_index file and you can build your in-memory name index. Reading the name_index file should give you *exactly* the same result as if you open every .meta file and build your in-memory name index. So, your job is to decide what you need in the name_index file so you can do the above. Same goes for the other two index files. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, April 14, 2007 7:48 am Subject: Re: Index To: cs551@merlot.usc.edu > Someone wrote: > > > In the spec we have the diagram: > > > > $(HomeDir) > > +- init_neighbor_list > > +- kwrd_index > > +- name_index > > +- sha1_index > > +- ... (other files you want to keep) > > +- files > > +- 1.data > > +- 1.meta > > +- 2.data > > +- 2.meta > > +- ... > > > > I am wondering if the three indexes should be logically > > separate, or just conceptually separate. > > You should have 3 physically separate files since the grading > guidelines calls for them. > > > For example, > > can we use a single struct to hold three indexes? > > In your program's memory, you can do whatever you want. But > you need to create 3 separate files when you "externalize" > what's in your program's memory. > > > In addition, > > in the above picture, what should we do to show the > > indexes are in the right place - write to three files (or > > a single file)? > > I'm not sure what you meant by "show the indices are in the right > place". They are shown as 3 separate files directly under > $(HomeDir) in the picture. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 07:56:37 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EEubo8004771 for ; Sat, 14 Apr 2007 07:56:37 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EEupPH032406 for ; Sat, 14 Apr 2007 07:56:51 -0700 Message-Id: <200704141456.l3EEupPH032406@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: get response message Date: Sat, 14 Apr 2007 07:56:51 -0700 From: william@bourbon.usc.edu Someone wrote: > If an intermediate node receives a get response that's > already expired (UOID not found in cache), should > the node still flip a coin to cache the file? Yes. The message cache is for routing/forwarding only. In general, it's probably a good idea to keep the networking related separate from the application logic. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 14 07:47:51 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3EElpOY003893 for ; Sat, 14 Apr 2007 07:47:51 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3EEm50x032243 for ; Sat, 14 Apr 2007 07:48:05 -0700 Message-Id: <200704141448.l3EEm50x032243@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Index Date: Sat, 14 Apr 2007 07:48:05 -0700 From: william@bourbon.usc.edu Someone wrote: > In the spec we have the diagram: > > $(HomeDir) > +- init_neighbor_list > +- kwrd_index > +- name_index > +- sha1_index > +- ... (other files you want to keep) > +- files > +- 1.data > +- 1.meta > +- 2.data > +- 2.meta > +- ... > > I am wondering if the three indexes should be logically > separate, or just conceptually separate. You should have 3 physically separate files since the grading guidelines calls for them. > For example, > can we use a single struct to hold three indexes? In your program's memory, you can do whatever you want. But you need to create 3 separate files when you "externalize" what's in your program's memory. > In addition, > in the above picture, what should we do to show the > indexes are in the right place - write to three files (or > a single file)? I'm not sure what you meant by "show the indices are in the right place". They are shown as 3 separate files directly under $(HomeDir) in the picture. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 22:39:53 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3E5drAs025447 for ; Fri, 13 Apr 2007 22:39:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3E5e8fY027132 for ; Fri, 13 Apr 2007 22:40:08 -0700 Message-Id: <200704140540.l3E5e8fY027132@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: doubt regarding deletion Date: Fri, 13 Apr 2007 22:40:08 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a doubt regarding the contents of the delete and the store message. > Specifically when a node sends a store message, in the certificate field do > we send the entire digitally signed filespec or the cert.pem file. In a STORE message, the "certification field" should contain the certificate file ("cert.pem"). > Which of these do we send in the delete. Digitally signed FileSpec. > My interpretation is that we send the > cert.pem of the originating node during store and during delete we send the > digitally signed filespec. Correct. > Also when a node receives a store , is it > supposed to store the certificate along with the .data and .mete files for > that particular file. Yes. You can call it .pem or .cert, and you should use the same file number as the .data and .meta files. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 19:26:31 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3E2QVen010222 for ; Fri, 13 Apr 2007 19:26:31 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3E2Qlqv024965 for ; Fri, 13 Apr 2007 19:26:47 -0700 Message-Id: <200704140226.l3E2Qlqv024965@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: File seq no and response to GET Date: Fri, 13 Apr 2007 19:26:47 -0700 From: william@bourbon.usc.edu Someone wrote: > I think these have been mentioned before, but I > want to make sure: > > 1. Do files in both cache and permanent space > share the same file number sequence? For > example, if node A stores a file in cache named > 1.data, 1.meta and 1.pem. After that node B > stores another file, this time in permanent space, > should this second file be numbered 2.data, 2.meta > and 2.pem, or still numbered 1 (if it's the first file > in the permanent space but the event occurs after > the first file in cache)? I would use share the same number sequence. This way, the 3 major indices will only need to point to a file number. Also, the difference between permanent and cache is fairly minor, so I would just do something extra to take care of the differences as oppose to do a lot of work to take care of the differences.. > 2. To respond to a search request, should both > cache and permanent space be searched, or > just searching the permanent space? Both. Again, there is very little difference between cache and permanent spaces. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 15:18:46 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3DMIkUj023270 for ; Fri, 13 Apr 2007 15:18:46 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3DMJ3M6022877 for ; Fri, 13 Apr 2007 15:19:03 -0700 Message-Id: <200704132219.l3DMJ3M6022877@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on DELETE Date: Fri, 13 Apr 2007 15:19:03 -0700 From: william@bourbon.usc.edu Someone wrote: > Say two nodes have the same copy of a file (filename, nonce, and > sha1), and one of them try to 'delete'. It first tries to delete the local > copy, however the signature verification fails. Should it stop or continue > flooding out Delete to other nodes? Yes. I guess it's possible that the local copy is somehow corrupted. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 15:03:24 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3DM3Osh022026 for ; Fri, 13 Apr 2007 15:03:24 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3DM3fIo022552 for ; Fri, 13 Apr 2007 15:03:41 -0700 Message-Id: <200704132203.l3DM3fIo022552@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Doubt on GET message Date: Fri, 13 Apr 2007 15:03:41 -0700 From: william@bourbon.usc.edu Someone wrote: > Given the uniqueness of FileID, if a node finds that he > has the file requested by GET request, is it OK for > this node to stop forwarding this GET request and > just send back the file through GET_R? In other words, > could we say that if there's a match, this will be the > only match in the network so we no longer need to > forward this GET (other nodes can't have this file)? To provide "plausible deniability", you should continue to flood. (This should also simply your code, slightly.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 09:36:56 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3DGaucI028534 for ; Fri, 13 Apr 2007 09:36:56 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3DGbEah018667 for ; Fri, 13 Apr 2007 09:37:14 -0700 Message-Id: <200704131637.l3DGbEah018667@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Test Cases Date: Fri, 13 Apr 2007 09:37:14 -0700 From: william@bourbon.usc.edu Someone wrote: > Can u please put back the link for the test cases web page?. > We need the test cases and the intructuions below it to get the > testcases and run the scripts to change the port no.s > automaticallyl. Oops! Sorry that I accidentically deleted it from the spec. It's put back now. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 13 07:37:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3DEb9vP019105 for ; Fri, 13 Apr 2007 07:37:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3DEbR69017406 for ; Fri, 13 Apr 2007 07:37:27 -0700 Message-Id: <200704131437.l3DEbR69017406@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Indices Date: Fri, 13 Apr 2007 07:37:27 -0700 From: william@bourbon.usc.edu Someone wrote: > Should we write out three indices to disk constantly to > preserve the info in case of program crash? You don't have to. But you can also do it as soon as they change. But this may generate too many writes. Another way is to trigger the write using a timer. When the timer expires, you can check if an index has been modified since the last write. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 21:54:07 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3D4s6jX005138 for ; Thu, 12 Apr 2007 21:54:06 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3D4sQJp008485 for ; Thu, 12 Apr 2007 21:54:26 -0700 Message-Id: <200704130454.l3D4sQJp008485@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Private and Public keys Date: Thu, 12 Apr 2007 21:54:26 -0700 From: william@bourbon.usc.edu Someone wrote: > I had a small question. Do we need to create the public and > private keys everytime a node restarts or Is it to be made (on > every node)only when the SERVANT network is booted for the first > time??? Every time your node starts, you should check if these files exist. If they do not exist, you should generate them. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 20:22:17 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3D3MHvt030303 for ; Thu, 12 Apr 2007 20:22:17 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3D3MbhI007201 for ; Thu, 12 Apr 2007 20:22:37 -0700 Message-Id: <200704130322.l3D3MbhI007201@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Part2 Date: Thu, 12 Apr 2007 20:22:37 -0700 From: william@bourbon.usc.edu Someone wrote: > Can we have sub directories under the files directory to > differentiate between permanent store area and cache area. Yes. It's up to you! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 11:43:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3CIhw7M020917 for ; Thu, 12 Apr 2007 11:43:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3CIgb2N000728 for ; Thu, 12 Apr 2007 11:42:37 -0700 Message-Id: <200704121842.l3CIgb2N000728@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About LRU Date: Thu, 12 Apr 2007 11:42:37 -0700 From: william@bourbon.usc.edu Hi, Just want to clarify... What I thought the question was about your index structure to keep track of LRU. That has no limit. Certainly, the total size of the cahced data file (you don't have to count metadata and certificate file sizes, and it's okay if you do) has a limit which is specified by the CacheSize key in the startup configuration file. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Sat, 07 Apr 2007 16:18:43 -0700 From: william@bourbon.usc.edu To: cs551@merlot.usc.edu Subject: Re: About LRU Someone wrote: > What is the size limit for cache used in implementing > LRU? (e.g. bytes or number of files). There is no limit. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 10:28:04 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHS4UN015012 for ; Thu, 12 Apr 2007 10:28:04 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHQhsX032080 for ; Thu, 12 Apr 2007 10:26:43 -0700 Message-Id: <200704121726.l3CHQhsX032080@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS530: Regarding Term Paper Date: Thu, 12 Apr 2007 10:26:43 -0700 From: william@bourbon.usc.edu Oops! Wrong class! Sorry! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Thu, 12 Apr 2007 10:26:01 -0700 From: william@bourbon.usc.edu To: cs551@merlot.usc.edu Subject: Re: CS530: Regarding Term Paper Someone wrote: > As i have not started working on my term paper, yet. But i am > quite sure that i would need to add some more reference as i am > working on it. So can we add the reference after we have > submitted the proposal. I hope that we will be able to add some > more papers. Sure! But you must submit 3 qualified references for approval by the end of today. If you do not submit anything, you will lose 20 out of 100 points. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 10:27:22 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHRMC1014851 for ; Thu, 12 Apr 2007 10:27:22 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHQ1xl032046 for ; Thu, 12 Apr 2007 10:26:01 -0700 Message-Id: <200704121726.l3CHQ1xl032046@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS530: Regarding Term Paper Date: Thu, 12 Apr 2007 10:26:01 -0700 From: william@bourbon.usc.edu Someone wrote: > As i have not started working on my term paper, yet. But i am > quite sure that i would need to add some more reference as i am > working on it. So can we add the reference after we have > submitted the proposal. I hope that we will be able to add some > more papers. Sure! But you must submit 3 qualified references for approval by the end of today. If you do not submit anything, you will lose 20 out of 100 points. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 10:20:25 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHKPjK014362 for ; Thu, 12 Apr 2007 10:20:25 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHJ32T031919 for ; Thu, 12 Apr 2007 10:19:03 -0700 Message-Id: <200704121719.l3CHJ32T031919@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg File Storage. Date: Thu, 12 Apr 2007 10:19:03 -0700 From: william@bourbon.usc.edu Someone wrote: > The same question with some changes... > 1) When there is a search message and when the nodes receives > responses... Suppose consider the case : > A-->B-->C-->D.A searches a file which is present in B,C,D so all > of them reply and suppose user(me) A choses the file from D. So > the file has to be sent from D to C and then from C to B and then > from B to A. So should B and C flip a coin and choose to store > it(they mite choose to store or not > > (If they choose to store then it will be 2 copies of the same > file in the same node with the same metadata.. is it ok?) No. It's not okay. You must not have more than one copy of the same file (identical Filename, Nonce, and SHA1). > ... but should they flip a coin?)? or they should not store, as > they already have the same file with same nonce and sha1?(and > hence there would be no need to flip a coin). Please see above. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 12 10:18:13 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHIDn5014186 for ; Thu, 12 Apr 2007 10:18:13 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3CHGqWi031874 for ; Thu, 12 Apr 2007 10:16:52 -0700 Message-Id: <200704121716.l3CHGqWi031874@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg File Storage Date: Thu, 12 Apr 2007 10:16:52 -0700 From: william@bourbon.usc.edu Someone wrote: > 1) When there is a search message and when the nodes receives > responses... Suppose consider the case : > A-->B-->C-->D. A searches a file which is > present in B,C,D so all of them reply and suppose user(me) A > choses the file from D. So the file has to be sent from D > to C and then from C to B and then from B to A. So should B and C > flip a coin and choose to store it(they mite choose to store or > not... but should they flip a coin?)? or they should not store, > as they already have the same file with same nonce and sha1?(and > hence there would be no need to flip a coin). B and C should flip a coin according to CacheProb. If they decide to store it, soon they should realize that they already have a copy. (If you want to check before flipping the coin, that would be fine too.) > 2) Should we store the no. of bytes used in the cache in a file, > so as to reload it when the nodes goes down and comes up again.? > or should we delete everythin in cache when the node goes > down(and so the no. of bytes will be 500KB,and there would be no > need to stored the no. of bytes used)? As I've mentioned before, you *must not* delete the cache when your node goes down. If you start the node with "-reset", then you should delete all files. I suggested that you may want to keep a LRU list. In this list, you can also store the file sizes. You can also make a similar list for permanent files. Such a list should be kept in a file just like the 3 main index files. > 3)I think for permanent we have to do this for sure(that is store > the no. of bytes used and reload the info again when it goes down > and starts up again)...is it right? Correct. You should do the same for cached files too. > 4)Whenever a file is being sent as a part of store request .. we > might not store it if while flippin a coin if it says "need not > store".. in that case as well should we flood it(regardless of > whether it was stored at that node)? And if we decide to flood > it, then wer should we store it ? in the ../homeDir/ or in cwd > (memory of which is not accounted for?) or in ../Files/? It should still flood to its neighbors according to NeighborStoreProb even if it decides not to cache this file. > 5)Can we have something like Cache,Permanent,Temporary > directories in the Files folder? This will make it all simpler.. > Files in Temporary might be deleted very soon.. becos if a node A > decides to forward store req(as a part of flooding) to B but B > has already recieved the store request via some other path... so > this is a replayed Message... but when node A sends the store > request B cant avoid it and B has to accept the file(should we > store it and delete it.. or not store it at all as we know that > it is a replayed message from the header and just ignore the > message comin and not store it.)? It's up to you, as long as all the cached and permanent files are under the "files" subdirectory. Personally, I think it's easier to put cached and permanent files in the same directory and put temporary files in /tmp. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 11 19:58:57 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3C2wv9r011036 for ; Wed, 11 Apr 2007 19:58:57 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3C2vctJ021855 for ; Wed, 11 Apr 2007 19:57:38 -0700 Message-Id: <200704120257.l3C2vctJ021855@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg File Storage Date: Wed, 11 Apr 2007 19:57:38 -0700 From: william@bourbon.usc.edu Someone wrote: > I would like to know whether my understanding is correct: > > A file is stored in Permanent Storage only when > 1)the Store Request originates at that node > 2)It gets a response to it's get message > > In all other cases the file is stored in Cache. Correct. > I have doubt as to where the file should be stored when Store > Request is flooded & and recieved at the node(so the origin of > the store is a different node)? Cache or in Permanent? If this node did not originate the STORE request, then according to the logic above, it should be stored in the cache. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 11 19:56:40 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3C2ue6E010866 for ; Wed, 11 Apr 2007 19:56:40 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3C2tLbb021808 for ; Wed, 11 Apr 2007 19:55:21 -0700 Message-Id: <200704120255.l3C2tLbb021808@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on GET Date: Wed, 11 Apr 2007 19:55:21 -0700 From: william@bourbon.usc.edu Someone wrote: > What should happen when a user is asked whether to replace the > file with the same filename in the current directory after 'get' > but user choose not to replace? Should the program: > > 1. Persist to create a file in current dir. but using 'filename' > from metadata. > Or > 2. Discard everything totally (ie. neither create a file in > current dir nor the permanent storage) Actually, it's neither. The copying of the file into the current working directory should be handled separatedly (after you have done saving the file in permanent space). At this point, you should just not copy the file into the current working directory. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 11 12:33:04 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3BJX4sx008129 for ; Wed, 11 Apr 2007 12:33:04 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3BJVkYB016044 for ; Wed, 11 Apr 2007 12:31:46 -0700 Message-Id: <200704111931.l3BJVkYB016044@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: doubt in search response Date: Wed, 11 Apr 2007 12:31:46 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a doubt regarding the search response. > > I wanted to ask how are supposed to handle the condition where we > get search responses from multiple nodes for a single request. Do > we indicate different responses with different set of serial > numbers or do we print all the responses contiguously. The example given in the spec is as follows: [1] FileID=02adefc1dfc97a082fa18a5ef1e8c487259b7fb4 FileName=foo FileSize=123 SHA1=b83a758fecbefcd3ea547fbf0f9a97eba0ea984c Nonce=01b7a1bd6f169dde22518a865ab2f44c70fcab82 Keywords=key1 key2 key3 [2] FileID=45929c03a7c84687a73543cc348484edc3829496 FileName=bar FileSize=4567 SHA1=6b6c5636c484d47599d20191c3023b8a29b2fe11 Nonce=fe1834fdf8cd7356ca11e0c77ac38d387e228f94 Keywords=key4 key5 [3] ... So, you need to keep a list of mapping between the item numbers (I guess you called these "serial numbers") you generated and their corresponding FileIDs. If the files are the same, you still need to assign a different item number. > Also do we > start to print out responses as they arrive or do we wait for a > certain timeout. You must start displaying them as soon as you get them. If the user types , you should terminate current search and return the prompt to the user so the user can enter the next command (which is expected to be a GET command). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 11 10:04:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3BH4IOL027895 for ; Wed, 11 Apr 2007 10:04:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3BH30tO014291 for ; Wed, 11 Apr 2007 10:03:00 -0700 Message-Id: <200704111703.l3BH30tO014291@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Final1 project regrading office hours and appointment times Date: Wed, 11 Apr 2007 10:03:00 -0700 From: william@bourbon.usc.edu Hi, Just to clarify... Between 10am and 1pm is for "walk-ins" (no appointments) *only*. So, if you would like an appointment, please let Pritam know when you'd like them for time not during the 3 walk-in hours. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Wed, 11 Apr 2007 00:58:26 -0700 From: Pritam Pradhan To: cs551@merlot.usc.edu Subject: Final1 project regrading office hours and appointment times Hi, The office hours for the regrading of the final1 project are going to be held on Friday 04/13 from 10:00am - 1:00pm. Please make appointments between 8:3 0am - 10am or between 1:30pm - 6:30pm on Friday 04/13(I encourage you to make an appointment). The available slots are placed every half an hour. If you are making an appointment please give 3 preferences of time. Also please mention in which of the following 3 categories you fall into: 1) If your score is 1 or very low, you should only make a 15-minute appointment to "get around the grading obstacle". 2) If not (1) and only asking to regrade one sub-section(as in (B)(1) ), also make a 15-minute appointment. 3) If neither (1) nor (2), make a 30-minute appointment. If you can't make it on Friday, please let me know. I may hold office hours on Monday 04/16 depending on whether there is any demand for the same. The room for office hours and appointments is SAL 211. Sincerely, Pritam Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 11 07:47:01 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3BEl1q5017029 for ; Wed, 11 Apr 2007 07:47:01 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3BEjis4012702 for ; Wed, 11 Apr 2007 07:45:44 -0700 Message-Id: <200704111445.l3BEjis4012702@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Part II beacon's init_neighbor_list Date: Wed, 11 Apr 2007 07:45:44 -0700 From: william@bourbon.usc.edu Someone wrote: > Would beacon nodes be provided with init_neighbor_list, > or they should build their own from ini file, in Part II? For a beacon node, it needs to find its neighbors from the startup configuration file and ignore init_neighbor_list just like in part (1). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 10 14:10:42 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3ALAgIX030388 for ; Tue, 10 Apr 2007 14:10:42 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3AL9Rav002044 for ; Tue, 10 Apr 2007 14:09:27 -0700 Message-Id: <200704102109.l3AL9Rav002044@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CS 551 Final Project Part (1) grade... Date: Tue, 10 Apr 2007 14:09:27 -0700 From: william@bourbon.usc.edu Hi, Just want to mention that since there are a lot of 1's, I expect the average to be higher after all the regrades are done. And by the way, please read the content of the previous email carefully! Thanks! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Tue, 10 Apr 2007 14:03:11 -0700 From: william@bourbon.usc.edu Subject: CS 551 Final Project Part (1) grade... ... The statistics for CS 551 Final Project Part (1) are: Count = 32 Avg = 57.67 StdDev = 44.11 Max = 114.50 Min = 1.00 2 110+ XX 3 105+ XXX 2 100+ XX 2 95+ XX 2 90+ XX 4 85+ XXXX 0 80+ 2 75+ XX 0 70+ 0 65+ 0 60+ 0 55+ 2 50+ XX 0 45+ 0 40+ 1 35+ X 1 30+ X 1 25+ X 0 20+ 0 15+ 0 10+ 0 5+ 10 0+ XXXXXXXXXX ... Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 10 10:56:02 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3AHu2Ve014339 for ; Tue, 10 Apr 2007 10:56:02 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3AHsm93031520 for ; Tue, 10 Apr 2007 10:54:48 -0700 Message-Id: <200704101754.l3AHsm93031520@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Delete Date: Tue, 10 Apr 2007 10:54:48 -0700 From: william@bourbon.usc.edu Someone wrote: > But if the verification fails, it means the DELETE message > is not issued by the node who originally stores the file. > Based on our rule no one should delete the file requested > to be deleted. Why should we continue forwarding this > DELETE request? Can you think of a case where this may not be true? -- Bill Cheng // bill.cheng@usc.edu On 4/10/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > If recipient of a DELETE message finds that > > the message fails verification, can this node > > stop forwarding this DELETE message? > > No! When a node sees a request message, it should simply > decrement the TTL and forwards the message to its neighbors. > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 10 10:40:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3AHe925013121 for ; Tue, 10 Apr 2007 10:40:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3AHctQB031218 for ; Tue, 10 Apr 2007 10:38:55 -0700 Message-Id: <200704101738.l3AHctQB031218@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Delete Date: Tue, 10 Apr 2007 10:38:55 -0700 From: william@bourbon.usc.edu Someone wrote: > If recipient of a DELETE message finds that > the message fails verification, can this node > stop forwarding this DELETE message? No! When a node sees a request message, it should simply decrement the TTL and forwards the message to its neighbors. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 9 21:42:31 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4gVk3016381 for ; Mon, 9 Apr 2007 21:42:31 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4fJNw022881 for ; Mon, 9 Apr 2007 21:41:19 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l3A4fJbj022880 for cs551@merlot; Mon, 9 Apr 2007 21:41:19 -0700 Date: Mon, 9 Apr 2007 21:41:19 -0700 From: william@bourbon.usc.edu Message-Id: <200704100441.l3A4fJbj022880@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Final project part (2) extension... Hi, Due to the unexpected power outage on the class web server, I have extended the final project part (2) deadline by two days to 11:45pm on Sunday, 4/22/2007. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 9 21:20:04 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4K4bg014523 for ; Mon, 9 Apr 2007 21:20:04 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4IqxH022583 for ; Mon, 9 Apr 2007 21:18:52 -0700 Message-Id: <200704100418.l3A4IqxH022583@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: More question on GET Date: Mon, 09 Apr 2007 21:18:52 -0700 From: william@bourbon.usc.edu Someone wrote: > When a node performs 'get', and the file happen to be in local storage: > > If the file is permanent, it should > 1. Make one copy in current directory. > 2. Do nothing to 'files' directory Correct. > If the file is cached, it should > 1. Make one copy in current directory. > 2. Make another copy in 'files' directory but stored in permanent > space > Or > Turn the cache copy to permanent copy and adjust storage size > accordingly. It should turn the cache copy to be a permanent copy and do whatever adjustments necessary. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 9 21:17:54 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4Hr2h014200 for ; Mon, 9 Apr 2007 21:17:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3A4GgnV022550 for ; Mon, 9 Apr 2007 21:16:42 -0700 Message-Id: <200704100416.l3A4GgnV022550@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Question on GET Date: Mon, 09 Apr 2007 21:16:42 -0700 From: william@bourbon.usc.edu Someone wrote: > When a node successfully gets a file but there is not enough space in > the permanent storage, it should discard the file and print error msg on the > console. However, there should always be one copy in the current directory > no matter what, correct? I think if there is no space in the permanent storage, you should *not* copy the file in the current working directory. The idea here is that "you may not get this file if you don't server this file". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 8 22:34:49 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l395YnKm003974 for ; Sun, 8 Apr 2007 22:34:49 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l395XeUN006055 for ; Sun, 8 Apr 2007 22:33:40 -0700 Message-Id: <200704090533.l395XeUN006055@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Log sha1 search Date: Sun, 08 Apr 2007 22:33:40 -0700 From: william@bourbon.usc.edu Someone wrote: > I am actually referring to: > > SHRQ > > If the search type if "sha1hash", the corresponding query field > would be the sha1 value, correct? Do we only log the last 4-byte of this or > the entire sha1? Oh! Yes, just the last 4 bytes please. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Sunday, April 08, 2007 8:34 PM To: cs551@merlot.usc.edu Subject: Re: Log sha1 search Someone wrote: > Do we need to log only the last 4 bytes of sha1 keyword (as with > or ) if the search type of the search request is > sha1? For search responses, in the data-dependent part, you only need to log the UOID of the search request message that this response is responding to. So, you don't need to log anything else. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 8 20:35:05 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l393Z5S1026908 for ; Sun, 8 Apr 2007 20:35:05 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l393XvED005036 for ; Sun, 8 Apr 2007 20:33:57 -0700 Message-Id: <200704090333.l393XvED005036@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Log sha1 search Date: Sun, 08 Apr 2007 20:33:57 -0700 From: william@bourbon.usc.edu Someone wrote: > Do we need to log only the last 4 bytes of sha1 keyword (as with > or ) if the search type of the search request is > sha1? For search responses, in the data-dependent part, you only need to log the UOID of the search request message that this response is responding to. So, you don't need to log anything else. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 8 11:11:52 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l38IBqQh014848 for ; Sun, 8 Apr 2007 11:11:52 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l38IAjlD031093 for ; Sun, 8 Apr 2007 11:10:45 -0700 Message-Id: <200704081810.l38IAjlD031093@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Problem with Merlot Date: Sun, 08 Apr 2007 11:10:45 -0700 From: william@bourbon.usc.edu Someone wrote: > I am unable to access the page on Merlot Server. I have seen this > problem since 2days. Its quite truoublesome as we need > to access the grading guidelines and documents periodically. There was an unexpected power outage last night around 7:30pm. Apparently, the machine in question did not survive a reboot. I went in and manually rebooted the machine this morning around 9am. Things should be fine now. The server was fine except for the duration mentioned above. If you have trouble connecting to the server at other times, please let me know as soon as possible! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 8 11:03:51 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l38I3pml014189 for ; Sun, 8 Apr 2007 11:03:51 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l38I2jAD030864 for ; Sun, 8 Apr 2007 11:02:45 -0700 Message-Id: <200704081802.l38I2jAD030864@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj part2 Date: Sun, 08 Apr 2007 11:02:45 -0700 From: william@bourbon.usc.edu Someone wrote: > If -reset option is given in the command line then should we > delete all the files in the files directory when the node > starts up. Yes! Everything! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 7 16:20:02 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l37NK2Vj011115 for ; Sat, 7 Apr 2007 16:20:02 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l37NIhox019630 for ; Sat, 7 Apr 2007 16:18:43 -0700 Message-Id: <200704072318.l37NIhox019630@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About LRU Date: Sat, 07 Apr 2007 16:18:43 -0700 From: william@bourbon.usc.edu Someone wrote: > What is the size limit for cache used in implementing > LRU? (e.g. bytes or number of files). There is no limit. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 7 15:36:59 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l37MawVR007537 for ; Sat, 7 Apr 2007 15:36:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l37MZeRh018876 for ; Sat, 7 Apr 2007 15:35:40 -0700 Message-Id: <200704072235.l37MZeRh018876@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Store probability Date: Sat, 07 Apr 2007 15:35:40 -0700 From: william@bourbon.usc.edu Someone wrote: > Thanks for the message. I still do not understand which > "*Prob" parameter would affect the *original sender* of STORE > in terms of whether he sends out the STORE request > to a particular neighbor. The spec implies that STORE > request is NOT sent to ALL neighbor by the node > that initiates the request (i.e. a neighbor is picked to receive > my request only probabilistically). Please correct me if I'm wrong > on this. The original sender must store the file! If there is no space in the permanent space, it does not flood at all. If it can store the file in its permanent space, it then uses NeighborStoreProb to probabilistically flood STORE messages to its neighbors. -- Bill Cheng // bill.cheng@usc.edu On 4/7/07, william@bourbon.usc.edu wrote: > Someone wrote: > > > The spec requires that "The store message should > > be probabilistically flooded to the whole network > > (up to a specified TTL)". > > > > However, the "StoreProb" parameter seems to deal > > only with a *receiving* node's decision to cache a copy > > or not. The other two "*Prob" parameters do not have > > to do with STORE message *sender*, either. Did I > > miss something here? > > "NeighborStoreProb" also affects probabilistic flooding of > STORE messages. When a node receives a STORE message, it > should flip a coin according to "StoreProb" to decide if it > should keep a copy of the file. Then, for *each of its > active neighbor* (except for the one the message came from), > it should flip a coin according to "NeighborStoreProb" to > decide if it should forward a copy of the STORE message to > this neighbor. > -- > Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 7 14:06:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l37L6YZw032704 for ; Sat, 7 Apr 2007 14:06:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l37L5GMq017744 for ; Sat, 7 Apr 2007 14:05:16 -0700 Message-Id: <200704072105.l37L5GMq017744@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Store probability Date: Sat, 07 Apr 2007 14:05:16 -0700 From: william@bourbon.usc.edu Someone wrote: > The spec requires that "The store message should > be probabilistically flooded to the whole network > (up to a specified TTL)". > > However, the "StoreProb" parameter seems to deal > only with a *receiving* node's decision to cache a copy > or not. The other two "*Prob" parameters do not have > to do with STORE message *sender*, either. Did I > miss something here? "NeighborStoreProb" also affects probabilistic flooding of STORE messages. When a node receives a STORE message, it should flip a coin according to "StoreProb" to decide if it should keep a copy of the file. Then, for *each of its active neighbor* (except for the one the message came from), it should flip a coin according to "NeighborStoreProb" to decide if it should forward a copy of the STORE message to this neighbor. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Apr 7 14:02:48 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l37L2mEr032379 for ; Sat, 7 Apr 2007 14:02:48 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l37L1T8R017692 for ; Sat, 7 Apr 2007 14:01:29 -0700 Message-Id: <200704072101.l37L1T8R017692@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: index files Date: Sat, 07 Apr 2007 14:01:29 -0700 From: william@bourbon.usc.edu Someone wrote: > Do we have to use binary tree structures for maintaining the > index memory structures? What if I am using STL containers? > Should I use one that employs binary search? There is no requirement what data structure you must use or what performance level you must have. You can use whatever you'd prefer. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 6 20:55:01 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l373t175012568 for ; Fri, 6 Apr 2007 20:55:01 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l373rjUP005825 for ; Fri, 6 Apr 2007 20:53:45 -0700 Message-Id: <200704070353.l373rjUP005825@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: TTL = 0 Date: Fri, 06 Apr 2007 20:53:45 -0700 From: william@bourbon.usc.edu Someone wrote: > Does the following mean that the status command should be ignored > completely? > > A value of zero means not to flood the status message ([BC: > Updated 4/4/2007] i.e., should only attempts to get status for the > node where the command was issued). No. It means that the node where you typed the command should respond to itself. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Apr 6 20:36:29 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l373aTFW011154 for ; Fri, 6 Apr 2007 20:36:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l373ZDDD005470 for ; Fri, 6 Apr 2007 20:35:13 -0700 Message-Id: <200704070335.l373ZDDD005470@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: *.pem file path Date: Fri, 06 Apr 2007 20:35:13 -0700 From: william@bourbon.usc.edu Someone wrote: > Are the *.pem files to be stored in the HomeDir path in the > startup.ini file? I assume you meant the public key certificate and the private key file. If that's the case, you should move them into somewhere in HomeDir. > Is the directory structure for storing permanent and cached > files totally upto our judgement? Yes. But you must name the files that are stored and retrieved ###.* where ### are numbers. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 5 23:05:50 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3665oOO006122 for ; Thu, 5 Apr 2007 23:05:50 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3664bfL025464 for ; Thu, 5 Apr 2007 23:04:37 -0700 Message-Id: <200704060604.l3664bfL025464@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Project 2 problem Date: Thu, 05 Apr 2007 23:04:37 -0700 From: william@bourbon.usc.edu Someone wrote: > I am having a peculiar problem. When i am copying the last 4 > bytes of SHA and MD5 hashes in an array. I have noticed that if > there is a 0 in the last 4 values then it doesn't get copied > > example: 70f7 gets copied as 7f7 > > As far as i know i am using the correct format > > example : sprintf(buff,"%01x""%01x",shabuf[18]); > > Every piece of values are working perfectly fine except if there > is a 0 in the last four hex values.What Should be the error.?? I'm not understand exactly what you did. To print a single byte of binary data as a 2-character hexstring, you should change your example to: sprintf(buff, "%02x", shabuf[18]); -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Apr 5 07:53:19 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l35ErJ3A029842 for ; Thu, 5 Apr 2007 07:53:19 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l35Eq85S015203 for ; Thu, 5 Apr 2007 07:52:08 -0700 Message-Id: <200704051452.l35Eq85S015203@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: mem bufsiz limit Date: Thu, 05 Apr 2007 07:52:08 -0700 From: william@bourbon.usc.edu Someone wrote: > Does the max memory buffer size limit applies > to sender as well (e.g. if sender sends a 8000K > file, it must malloc() 1000 times)? You should only use a single buffer for this! Once you see that it's a GET response, you should create a temporary file and start writing the content of the buffer into this temporary file. When you are done, you can move (and rename) the file into the "files" directory. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 4 22:07:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3557XIj015640 for ; Wed, 4 Apr 2007 22:07:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3556NjH009192 for ; Wed, 4 Apr 2007 22:06:23 -0700 Message-Id: <200704050506.l3556NjH009192@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Suppress openssl stdout output Date: Wed, 04 Apr 2007 22:06:23 -0700 From: william@bourbon.usc.edu Someone wrote: > How could we suppress openssl's output on stdout when > it creates the two keys? I tried ">/dev/null" but it still > writes to the console. There is no need to suppress its output! I would suggest you do this before your sv_node goes multithread! If you really want to suppress it, I think you can take the "bash -c ..." approach mentioned in the spec. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 4 21:53:40 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l354reC3014392 for ; Wed, 4 Apr 2007 21:53:40 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l354qVax008840 for ; Wed, 4 Apr 2007 21:52:31 -0700 Message-Id: <200704050452.l354qVax008840@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About Bit-Vector Date: Wed, 04 Apr 2007 21:52:31 -0700 From: william@bourbon.usc.edu Someone wrote: > As the specs says "*Bit-vector* - hexstring-encoded (*256* characters long)" > but the bit-vector is only 1024 bits long that is 128 byte (128 chars) > so how should we show the 1024 bits in 256 chars, ie. how should we > divide the 128 bytes into 256 bytes? In hexstring encoding, a byte is displayed as 2 ASCII characters where the only allowed ASCII characters are '0' through '9' and 'a' through 'f'. For example, the hexstring representation of the byte 0x5a is "5a". So, if there are 128 bytes data, its hexstring representation will have 256 characters (plus a trailing '\0'). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Apr 4 21:43:41 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l354hf8F013561 for ; Wed, 4 Apr 2007 21:43:41 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l354gWEx008637 for ; Wed, 4 Apr 2007 21:42:32 -0700 Message-Id: <200704050442.l354gWEx008637@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: fileid, sha1, nonce Date: Wed, 04 Apr 2007 21:42:32 -0700 From: william@bourbon.usc.edu Someone wrote: > Can you define and their purpose > FileId > Nonce > Sha1 Please read the following *carefully* (and please read it yourself and don't rely on others to tell you what's in it): http://merlot.usc.edu/cs551-s07/projects/final.html#nonce SHA1 is just a function that an map any data to a 20-byte value. > Also, what should be the file content when u r creating a > digitasl signature for the file I'm not sure what you are asking. The only place we do a digital signature is to sign a File Spec for a DELETE message. A File Spec is exactly 3 lines long: FileName=... SHA1=... Nonce=... If you write a File Spec into a file, you can run the "openssl mime ..." command mentioned in: http://merlot.usc.edu/cs551-s07/projects/final.html#cert to create a digital signature for it and put the output of this command in a DELETE message. -- Bill Cheng // bill.cheng@usc.edu Return-Path: cs551@bourbon.usc.edu Delivery-Date: Wed Apr 4 13:07:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l34K7w45005543 for ; Wed, 4 Apr 2007 13:07:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l34K6ogL000789 for ; Wed, 4 Apr 2007 13:06:50 -0700 Received: (from cs551@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l34K6oMH000788 for cs551@merlot.usc.edu; Wed, 4 Apr 2007 13:06:50 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l34K6o0a000785 for ; Wed, 4 Apr 2007 13:06:50 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l34K6ol2000784 for cs551; Wed, 4 Apr 2007 13:06:50 -0700 Date: Wed, 4 Apr 2007 13:06:50 -0700 From: william@bourbon.usc.edu Message-Id: <200704042006.l34K6ol2000784@bourbon.usc.edu> To: cs551@bourbon.usc.edu Subject: clarification to the project spec... Hi, I've added a few clarifications to the project spec: 1) Item (6) in the Summary of Basic Message Types section: [BC: Updated 4/4/2007] If there is not enough space in the permanent area, this node should not flood store messages. 2) Item (7) in the Summary of Basic Message Types section: [BC: Updated 4/4/2007] If a node generates a search response message for a file in its cache, it should update the LRU order of this file. 3) "status" command in the Commandline Interface section: A value of zero means not to flood the status message ([BC: Updated 4/4/2007] i.e., should only attempts to get status for the node where the command was issued). 4) "store" command in the Commandline Interface section: A value of zero means not to flood the store message ([BC: Updated 4/4/2007] i.e., should only attempts to store the file in the node on which the command was issued). Regarding LRU, the spec does not mention other occasions when a cached file needs to move its LRU order since they were obvious. Please remember that point of LRU is that if a file is popular (because it was STOREd or SEARCHed), it should be moved to the end of the LRU chain to keep it around longer. So, for example, when a node decides to cache a file, it should put this file at the end of the LRU chain. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 3 22:44:04 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l345i4t6001845 for ; Tue, 3 Apr 2007 22:44:04 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l345gvdr022751 for ; Tue, 3 Apr 2007 22:42:57 -0700 Message-Id: <200704040542.l345gvdr022751@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Cache and permanent area Date: Tue, 03 Apr 2007 22:42:57 -0700 From: william@bourbon.usc.edu Someone wrote: > About what you described below: > > > I would put them all in the "files" subdirectory of the > > node's home directory. I would also have a LRU list to keep > > track of the URL order of cached files. Then I might as well > > have another list to keep track of the permanent files. > > These 2 lists need to be written to the disk when my node > > goes down (just like the other 3 index files). Also, for > > each file i, I would have i.cached and have a 0 or 1 in it to > > indicate if file i is permanent or cached. Anyway, this is > > just how I would do it. It is also possible that I may have > > missed something since this is off the top of my head right > > now. > > The "real" cache does not remember data when a computer > crashes or reboots (not like a disk) that is another > difference between temporary and permanent storage. So why do > we have to do that for our project? I hope that you see what > I meant. Because that's not the kind of caching we are talking about! We are doing something like a P2P cache, DNS cache or a web browser cache. When you kill IE or Firefox, everything in the cache don't disappear. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 3 21:56:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l344uX6h030084 for ; Tue, 3 Apr 2007 21:56:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l344tRLl021902 for ; Tue, 3 Apr 2007 21:55:27 -0700 Message-Id: <200704040455.l344tRLl021902@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Cache and permanent area Date: Tue, 03 Apr 2007 21:55:27 -0700 From: william@bourbon.usc.edu Someone wrote: > Should we use the same structure for in-memory cache > and permanent area in Part II ? I'm not sure what you are referring to. If you are thinking about keeping cached files in one directory and keeping permanent files in a different directory, I don't think it's a good idea. This is because if you do that, when you need to refer to a file, you need to say if it's in the cache or not. I would put them all in the "files" subdirectory of the node's home directory. I would also have a LRU list to keep track of the URL order of cached files. Then I might as well have another list to keep track of the permanent files. These 2 lists need to be written to the disk when my node goes down (just like the other 3 index files). Also, for each file i, I would have i.cached and have a 0 or 1 in it to indicate if file i is permanent or cached. Anyway, this is just how I would do it. It is also possible that I may have missed something since this is off the top of my head right now. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Tue Apr 3 15:17:59 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l33MHxoK030752 for ; Tue, 3 Apr 2007 15:17:59 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l33MGspf017375 for ; Tue, 3 Apr 2007 15:16:54 -0700 Message-Id: <200704032216.l33MGspf017375@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About sending file Date: Tue, 03 Apr 2007 15:16:54 -0700 From: william@bourbon.usc.edu Someone wrote: > Is there a limit to packet size? I assume you mean "message size". The limit is close to 2^32 since the Data Length field is a 32-bit number. > if it is there how we should send the file bigger than the packet size? 2^32 is large enough so that you don't have to worry about it. > (In store packet as well as get response packet) > can we use reserved byte for file sending purpose? No. But you should know how to do this already. You've done it in warmup project #1. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Mon Apr 2 22:31:49 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l335Vnca015573 for ; Mon, 2 Apr 2007 22:31:49 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l335Ulsi005423 for ; Mon, 2 Apr 2007 22:30:47 -0700 Message-Id: <200704030530.l335Ulsi005423@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: reg FileID Date: Mon, 02 Apr 2007 22:30:47 -0700 From: william@bourbon.usc.edu Someone wrote: > How should we create the FileID? You can call GetUOID(). > Is it a 20 byte random no. or a SHA1 hash of anything? > If its a SHA1 hash of something, wat is it? It's a 20-byte random string. And the way we are doing it in GetUOID(), it's a hash of something. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 22:46:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l325kveD030594 for ; Sun, 1 Apr 2007 22:46:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l325jx6F019952 for ; Sun, 1 Apr 2007 22:45:59 -0700 Message-Id: <200704020545.l325jx6F019952@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Nonce. Date: Sun, 01 Apr 2007 22:45:59 -0700 From: william@bourbon.usc.edu Someone wrote: > Im sorry for the confusion in the last mail. In this mail i would > refer everything correctly. > > > Not just the same Nonce, the entire File Metadata will be the > > same. If you do a STORE for a file together with its metadata, > > every copy of this file in the network will have exactly the > > same file data and file metadata. > > If the NONCE and SHA1 and other details were the same in all the > copies in the network, then when we get a get message (which > contains just the NONCE or FILE ID) which node will respond to > it? (as many nodes have the same file with the same metadata, > which was created when the store message was sent). A GET message only has a FileID. There is no NONCE in a GET message. Please read "Nonce & FileID" in the Some Definitions section of the spec. > And hence what i think is there is no difference between NONCE > and SHA1(as both are same for a certain file) and only in the > case of the same file being stored again the SHA1 changes. > Pls explain the difference between NONCE and SHA1(apart from the > thing that we will create the nonce with "nonce" and the SHA1 > with the "file"). NONCE is a random number. SHA1 is the content hash of the file. If you enter a STORE command 5 times at a node for the same file, the SHA1 values will stay the same but the NONCE will be different every time. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Sunday, April 1, 2007 7:02 pm Subject: Re: Reg Nonce. To: cs551@merlot.usc.edu > Someone wrote: > > > I have a doubt in the definition of the NONCE. > > In your message, I think when you said SHA1, you really meant > NONCE. The way we create a NONCE is to use GetUOID() to > generate a nonce that's the same size as SHA1. > > > In case of different nodes as the store message is forwarded the > > same metadata gets forwarded.So there is a chance of same SHA1 > > value being present for one file but in different nodes. > > Not just the same Nonce, the entire File Metadata will be the > same. If you do a STORE for a file together with its metadata, > every copy of this file in the network will have exactly the > same file data and file metadata. > > > How would there be a same file created in the same node with same > > SHA1 > > (is it becos this file was flooded and it reached again?... but > > in that case we would have ignored it with the help of UOID). > > I'm not sure which UOID are you referring to. > > > Why im asking this is because > > 1)the sequence_no which is given as input to the SHA function. > > And this sequence_no increases everytime and so if u create a > > file and delete it and create it again then > > the SHA1 value will be different. > > Exactly. The NONCE would be different. You don't even have > to delete it. If on the same node you issue a STORE command > twice with exactly the same set of keywords, they will be > considered *different* files since the File Meatadata are > different (because the NONCEs are different). > > > 2)And if the same file was created somewer else also the SHA1 > > would be different as the node_instance_id would change. > > Correct. > > > And in the case if deletion of a file , and if the store > > message(which was flooded when the file was created) comes to the > > origin then shud we ignore the message or create the message > > again(In the spec in the case of deleting it says you shud not > > delete the new copies if that node was the origin.. so i guess > > applying same rule leads us to not creating again). > > If you issue a DELETE command, all instances of the same file > (having the same FILENAME, SHA1, and NONCE) should be deleted. > The node that originates the DELETE command should delete this > file first if it has it. > > > Should we use "soda" as the object type so as to create the > > nonce(or the fileID.. i guess both are same... arent they?)? > > You can use anything (I would use "nonce"). > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 22:12:12 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l325CC2j027932 for ; Sun, 1 Apr 2007 22:12:12 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l325BDx1019453 for ; Sun, 1 Apr 2007 22:11:13 -0700 Message-Id: <200704020511.l325BDx1019453@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Part2 Date: Sun, 01 Apr 2007 22:11:13 -0700 From: william@bourbon.usc.edu Someone wrote: > This is what i get... > The bit vector for "blondie" is as follows > > 000000000000000000000000000000000000000000000000000000000 > 000000000000000000000000000000000000000000000000000800000 > 000000000000000000000000000000000000000000000000000000000 > 000000000000000000000000000000000000000000000000000000000 > 0000000000000000000080000000 > > > The bit vector for "title" is > > 0000000000000000000000000000000000000000000000000000000000 > 0000000000000000000000000001000000000000000000000000000000 > 0000000000000000000000000000000000000000000000000000000000 > 0000000000000000000000000000000000000000000000000000000000 > 000000000000000004000000 > > For all other keywords we get the right thing. Looks like your upper 512 bits are correct, but the lower 512 bits are wrong. So, let's see what's going on. The lower 512 bits come from MD5. Let's look at "blondie" first: echo -n "blondie" | openssl md5 returns "d05560f6566db54140c1c5adf58cf70f". The right 9 bits is "10f". Looks like you are turning on bit "1f" instead of bit "10f". Let's also look at "title": echo -n "title" | openssl md5 returns "d5d3db1765287eef77d7927cc956f50a". The right 9 bits is "10a". Looks like you are turning on bit "1a" instead of bit "10a". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 19:08:08 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l32288U0013520 for ; Sun, 1 Apr 2007 19:08:08 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3227AmB017577 for ; Sun, 1 Apr 2007 19:07:10 -0700 Message-Id: <200704020207.l3227AmB017577@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Part 2 Date: Sun, 01 Apr 2007 19:07:10 -0700 From: william@bourbon.usc.edu Someone wrote: > We were checking if the way we generate the bit-vector was > correct and while comparing it with the additional bit-vector > examples provided on the class webpage we found that all the bit > vectors were matching with ours except two of them namely the bit > vectors for the keywords "title" and "blondie". We want to know > if the bit vectors on the class webpage for those two keywords > are correct. The bit-vector for "title" is: 0000000000000000000000000000000000000000000000000000000000000000 0000000000000000000001000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000000000000000000400 0000000000000000000000000000000000000000000000000000000000000000 The bit-vector for "blondie" is: 0000000000000000000000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000080000000000000000000 0000000000000000000000000000000000000000000000000000000000008000 0000000000000000000000000000000000000000000000000000000000000000 If you are getting something different, please let me know exactly what you get. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 19:01:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l3221YF0012900 for ; Sun, 1 Apr 2007 19:01:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l3220a1U017345 for ; Sun, 1 Apr 2007 19:00:36 -0700 Message-Id: <200704020200.l3220a1U017345@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Reg Nonce. Date: Sun, 01 Apr 2007 19:00:36 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a doubt in the definition of the NONCE. In your message, I think when you said SHA1, you really meant NONCE. The way we create a NONCE is to use GetUOID() to generate a nonce that's the same size as SHA1. > In case of different nodes as the store message is forwarded the > same metadata gets forwarded.So there is a chance of same SHA1 > value being present for one file but in different nodes. Not just the same Nonce, the entire File Metadata will be the same. If you do a STORE for a file together with its metadata, every copy of this file in the network will have exactly the same file data and file metadata. > How would there be a same file created in the same node with same > SHA1 > (is it becos this file was flooded and it reached again?... but > in that case we would have ignored it with the help of UOID). I'm not sure which UOID are you referring to. > Why im asking this is because > 1)the sequence_no which is given as input to the SHA function. > And this sequence_no increases everytime and so if u create a > file and delete it and create it again then > the SHA1 value will be different. Exactly. The NONCE would be different. You don't even have to delete it. If on the same node you issue a STORE command twice with exactly the same set of keywords, they will be considered *different* files since the File Meatadata are different (because the NONCEs are different). > 2)And if the same file was created somewer else also the SHA1 > would be different as the node_instance_id would change. Correct. > And in the case if deletion of a file , and if the store > message(which was flooded when the file was created) comes to the > origin then shud we ignore the message or create the message > again(In the spec in the case of deleting it says you shud not > delete the new copies if that node was the origin.. so i guess > applying same rule leads us to not creating again). If you issue a DELETE command, all instances of the same file (having the same FILENAME, SHA1, and NONCE) should be deleted. The node that originates the DELETE command should delete this file first if it has it. > Should we use "soda" as the object type so as to create the > nonce(or the fileID.. i guess both are same... arent they?)? You can use anything (I would use "nonce"). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 18:52:10 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l321qAaG012238 for ; Sun, 1 Apr 2007 18:52:10 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l321pCRc017223 for ; Sun, 1 Apr 2007 18:51:12 -0700 Message-Id: <200704020151.l321pCRc017223@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About digital signature Date: Sun, 01 Apr 2007 18:51:12 -0700 From: william@bourbon.usc.edu Someone wrote: > 1. with the store packet we are sending the public key, which will be > used to verify the delete command, > so do we have to create these both (public and private) keys exactly > once, store them and use them for signing all the data originating on > the node (delete packet data)? Yes. When your node comes up, you should check if the public key certificate and the private key files exist. If they do not exist, you need to generate them. When you start your node with -reset, you should delete these files first. > 2. whats the difference between cache storage and permanent storage space? Cache is subject to LRU replacement and permanent storage is not. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 11:20:30 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l31IKTOo009491 for ; Sun, 1 Apr 2007 11:20:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l31IJWew013193 for ; Sun, 1 Apr 2007 11:19:32 -0700 Message-Id: <200704011819.l31IJWew013193@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About neighbor list in Part II Date: Sun, 01 Apr 2007 11:19:32 -0700 From: william@bourbon.usc.edu Someone wrote: > Since Part II does not require JOIN, how should the > program respond if it cannot find a neighbor list? If the node is a regular node and the init_neighbor_list is not present in its home directory, you can do what you were doingn in part (1) if you are confident that your JOIN code is working correctly. If you are not sure your JOIN code is working correctly, for part (2), you are allowed to just print an error message and quit. > Is it required to feed two things to the program > (startup file and neighbor list) for it to normally start, > or we could hard-code something (e.g. look for file > init_neighbor_list in home directory)? For a beacon node, it only needs the startup configuration file. For a regular nonde, it needs a startup configuration file and a init_neighbor_list file. We will make sure that a regular node will not start without a valid init_neighbor_list file. And we will make sure that all startup configuration file will have NoCheck set to 1. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sun Apr 1 11:16:39 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l31IGd7g009164 for ; Sun, 1 Apr 2007 11:16:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l31IFgxJ013133 for ; Sun, 1 Apr 2007 11:15:42 -0700 Message-Id: <200704011815.l31IFgxJ013133@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About bit vector Date: Sun, 01 Apr 2007 11:15:42 -0700 From: william@bourbon.usc.edu Someone wrote: > what are the contains of the vector which we are going to store > in the file? > are all the keywords only or filename also? Just the list of keywords. But please remember that you first have to convert everything to lowercase. > and the bit_vector you have given on the page, should our vector > match with that vector for the same case? It needs to match *exactly*. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 31 21:48:28 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l314mSGl012220 for ; Sat, 31 Mar 2007 21:48:28 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l314lXFP011275 for ; Sat, 31 Mar 2007 21:47:33 -0700 Message-Id: <200704010447.l314lXFP011275@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About bit vector Date: Sat, 31 Mar 2007 21:47:33 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a doubt, > as the specs says the bit vector is used while searching only, > and the search is carried out by keyword matching and not the bit-vector > matching, Keyword matching is very slow since it will require a lot of string comparisons. An efficient data structure for keyword matching is the Bloom Filter. Our bit-vector is kind of a simplified Bloom Filter. > then why do we need to store the bit vector in the meta data file? Since the bit-vector is derived from the list of keywords, it is redundant to have both keywords and bit-vector in the file metadata. So, the main reason that bit-vector is part of file metadata is for ease of grading, i.e., to see if you have calculated the bit-vector correctly. Therefore, you must put bit-vector in a file metadata. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 31 08:00:48 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2VF0dZe012174; Sat, 31 Mar 2007 08:00:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2VExkPT004425; Sat, 31 Mar 2007 07:59:46 -0700 Message-Id: <200703311459.l2VExkPT004425@bourbon.usc.edu> To: cs551@merlot.usc.edu, cs530@merlot.usc.edu, cs558l@merlot.usc.edu Subject: Re: class web server downtime Saturday morning... Date: Sat, 31 Mar 2007 07:59:46 -0700 From: william@bourbon.usc.edu Hi, The servers apparently automatically came on around 6am when the power came back on. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Fri, 30 Mar 2007 19:21:47 -0700 From: william@bourbon.usc.edu To: cs551@merlot.usc.edu, cs530@merlot.usc.edu, cs558l@merlot.usc.edu Subject: class web server downtime Saturday morning... Hi, I got the message below about another power shutdown in SAL this Saturday morning between 3:45am and 6:30am. So, I will be shutting down the class web server (merlot) and my e-mail server (bourbon) tonight probably around 10:30pm. So, if you would like to work on something that requires information on the class web server, please make a copy of the web pages. Hopefully, bourbon and merlot will automatically reboot when SAL gets power back around 6:30am tomorrow. If not, I have to come to campus to turn on these servers manually, and that may not happen until Monday. So, it's important to keep copies of web pages if your work this weekend depends on them. Sorry about the short notice. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Fri, 30 Mar 2007 11:40:14 -0700 From: CS Web Consultant To: cswebmst@usc.edu Subject: IT Alert! Prepare for power shutdown- SAL, PHE, WAH -31st March 20 07 Dear All, FMS will shutdown power in SAL, WAH and PHE between 3:45am and 6:30am on 31st March 2007 (tomorrow). Schedule and details are given in the attached notices. It is recommended that all computers and electrical equipment are shut down to prepare for this scheduled power outage. Thank you, Regards, Vishal Thakkar IT and Web Consultant Computer Science Department, USC SAL and WAH 4:15am to 6:15am PHE 3:45am to 6:30am Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 30 22:57:26 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2V5vQdC021108 for ; Fri, 30 Mar 2007 22:57:26 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2V5uXdi002803 for ; Fri, 30 Mar 2007 22:56:33 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l2V5uXPP002802 for cs551@merlot; Fri, 30 Mar 2007 22:56:33 -0700 Date: Fri, 30 Mar 2007 22:56:33 -0700 From: william@bourbon.usc.edu Message-Id: <200703310556.l2V5uXPP002802@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: HW2 deadline extension... Hi, Since there will be a power outage early tomorrow morning, I need to shutdown the submission server tonight before I go to sleep (and I'm going to sleep soon). So, I will extend HW2 submission deadline by 24 hours. Sorry about the last minute change, but I was only made aware of the power outage today. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 30 19:22:49 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2V2Mexi003770; Fri, 30 Mar 2007 19:22:40 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2V2LlMX032275; Fri, 30 Mar 2007 19:21:47 -0700 Message-Id: <200703310221.l2V2LlMX032275@bourbon.usc.edu> To: cs551@merlot.usc.edu, cs530@merlot.usc.edu, cs558l@merlot.usc.edu Subject: class web server downtime Saturday morning... Date: Fri, 30 Mar 2007 19:21:47 -0700 From: william@bourbon.usc.edu Hi, I got the message below about another power shutdown in SAL this Saturday morning between 3:45am and 6:30am. So, I will be shutting down the class web server (merlot) and my e-mail server (bourbon) tonight probably around 10:30pm. So, if you would like to work on something that requires information on the class web server, please make a copy of the web pages. Hopefully, bourbon and merlot will automatically reboot when SAL gets power back around 6:30am tomorrow. If not, I have to come to campus to turn on these servers manually, and that may not happen until Monday. So, it's important to keep copies of web pages if your work this weekend depends on them. Sorry about the short notice. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Fri, 30 Mar 2007 11:40:14 -0700 From: CS Web Consultant To: cswebmst@usc.edu Subject: IT Alert! Prepare for power shutdown- SAL, PHE, WAH -31st March 2007 Dear All, FMS will shutdown power in SAL, WAH and PHE between 3:45am and 6:30am on 31st March 2007 (tomorrow). Schedule and details are given in the attached notices. It is recommended that all computers and electrical equipment are shut down to prepare for this scheduled power outage. Thank you, Regards, Vishal Thakkar IT and Web Consultant Computer Science Department, USC SAL and WAH 4:15am to 6:15am PHE 3:45am to 6:30am Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 29 10:52:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2THq9he010262 for ; Thu, 29 Mar 2007 10:52:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2THpMCl001700 for ; Thu, 29 Mar 2007 10:51:22 -0700 Message-Id: <200703291751.l2THpMCl001700@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: final project part 2 bitvector in store msg Date: Thu, 29 Mar 2007 10:51:22 -0700 From: william@bourbon.usc.edu Someone wrote: > When sending the file metadata in the store command is the > bitvector also sent along or is it computed at the new node > anew? All fields in the File Metadata are *required*. They are: FileName FileSize SHA1 Nonce Keywords Bit-vector You send the whole File Metadata for messages that has a Metadata field. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 29 10:49:12 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2THnC6G009930 for ; Thu, 29 Mar 2007 10:49:12 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2THmPU4001650 for ; Thu, 29 Mar 2007 10:48:25 -0700 Message-Id: <200703291748.l2THmPU4001650@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: final project part 2: file meta data Date: Thu, 29 Mar 2007 10:48:25 -0700 From: william@bourbon.usc.edu Someone wrote: > The spec says: > The tag names are arbitrary. Both tag names and values must be > treated as searchable keywords > > So in the example, > store blondie1.mp3 30 \ > categories="audio mp3" \ > artist="Blondie" \ > title="Heart of Glass" \ > url="http://www.blondie.net/" \ > additional_keywords="debra harry" > > The question then: > So the tags can have any useful name and should be included in > the search? I assume they would be searched independantly of the > tag's value as each '=' or space separated word should have a > corresponding bit-vector. But does that make for useful search > results? If someone searches for keyword 'categories' then any > file, irrespective of the value of categories tag, having the > word categories would come up as a hit. Is that what the spec is > asking us to do? Yes. The reason is that people will not agree on what's on the left hand side of the equal side. For example, someone may use "band" instead of "artist". Since this is for some regular folks to use and this is entered in the commandline, it's probably better to just allow anything. The downside is that you can get things you don't want. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Mar 28 23:33:16 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2T6XGiX021111 for ; Wed, 28 Mar 2007 23:33:16 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2T6WUIr026585 for ; Wed, 28 Mar 2007 23:32:30 -0700 Message-Id: <200703290632.l2T6WUIr026585@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: hw2 doubt Date: Wed, 28 Mar 2007 23:32:30 -0700 From: william@bourbon.usc.edu Hi, I've added a link in the spec to what Leslie has written previously for easy access. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- Date: Wed, 28 Mar 2007 22:32:29 -0700 From: Leslie Cheung To: cs551@merlot.usc.edu Subject: Re: hw2 doubt Someone wrote: > In homework#2, after I am done with writing tcl script what is the file format > thar xgraph takes as input and how do I generate that file format to generate > the plot. You can check the ns2 tutorial at http://www.isi.edu/nsnam/ns/tutorial/ As for xgraph format, please see my email with timestamp "Thu 22 Feb 14:32". --Leslie Return-Path: william@bourbon.usc.edu Delivery-Date: Wed Mar 28 10:02:52 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2SH2qBs021970 for ; Wed, 28 Mar 2007 10:02:52 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2SH28U1015361 for ; Wed, 28 Mar 2007 10:02:08 -0700 Message-Id: <200703281702.l2SH28U1015361@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: About reseach paper Date: Wed, 28 Mar 2007 10:02:08 -0700 From: william@bourbon.usc.edu Someone wrote: > I am not able to get access to one of the research papers in the reading > list. When I tried opening a paper on BitTorrent p2p system > (http://merlot.usc.edu/cs551-s07/papers/Qiu04a.pdf), your web server > gave me following error. > > "You don't have permission to access /cs551-s07/papers/Qiu04a.pdf on > this server." > > Please help me out with this. Oops! It is fixed now. Thanks for letting me know! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 22:43:38 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P5hc4R016617 for ; Sat, 24 Mar 2007 22:43:38 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P5eD9n003511 for ; Sat, 24 Mar 2007 22:40:13 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l2P5eDgp003510 for cs551@merlot; Sat, 24 Mar 2007 22:40:13 -0700 Date: Sat, 24 Mar 2007 22:40:13 -0700 From: william@bourbon.usc.edu Message-Id: <200703250540.l2P5eDgp003510@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: submission for group project Hi, If you are doing the project with another student, only one student needs to submit (both would be fine too). By default, I will collect the last on-time submission from *each* student. I will sort them out later. Please read the output of bsubmit. If you see 2 or more "Verification Successful", then your submission made it to the server. If you want to view the submission ticket, please follow the information in the bsubmit output. (I'm going to sleep now.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 21:56:39 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4udpn012850 for ; Sat, 24 Mar 2007 21:56:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4rFvD002693 for ; Sat, 24 Mar 2007 21:53:15 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l2P4rFmn002691 for cs551@merlot; Sat, 24 Mar 2007 21:53:15 -0700 Date: Sat, 24 Mar 2007 21:53:15 -0700 From: william@bourbon.usc.edu Message-Id: <200703250453.l2P4rFmn002691@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: submission deadline approaching... Hi, It's two hours away from the submission deadline of final project part (1) and only a small number of students have made a submission. Please remember that the penalty is very harsh for late submissions. I strongly suggest that, if you have something that you can submit, make a submission *now*. If you improve it later, make another submission. If you have trouble getting your submission to work and it's around the deadline, I will not be able to give you special treatment. So, please *practice* a submission soon. Thanks! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 21:51:53 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4prQh012369 for ; Sat, 24 Mar 2007 21:51:53 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4mTrU002569 for ; Sat, 24 Mar 2007 21:48:29 -0700 Message-Id: <200703250448.l2P4mTrU002569@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Logging messages Date: Sat, 24 Mar 2007 21:48:29 -0700 From: william@bourbon.usc.edu Someone wrote: > I'm a little unclear about the difference for printing messages with > "s", "r", and "f". Please tell me if this is correct: > > - log a message with "s" if we are the originating node that is > sending it (i.e. we're not just flooding or routing a message that we > received) > - log a message with "f" if we got the message from someone else and > are either flooding or routing it > - log a message with "r" if we got the message that we don't forward > (flood or route) Exactly. > How important is it to have the correct labeling of message, as long > as the behavior (i.e. messages sent and received) is correct? It's not that important. It's good for debugging. It should not be too hard to distinguish the 3 cases. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 21:50:16 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4oG3j012351 for ; Sat, 24 Mar 2007 21:50:16 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4kpp0002534 for ; Sat, 24 Mar 2007 21:46:51 -0700 Message-Id: <200703250446.l2P4kpp0002534@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 21:46:51 -0700 From: william@bourbon.usc.edu Someone wrote: > sorry but one last question about check. Do we let the user > execute a command (shutdown or status) if check message was sent > and no response has been received or the timer for check hasnt > expired? Yes. CHECK is kind of a background activity. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 4:52 pm Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > if ( NOTIFY recvd on active connection || REMOTE SIDE TCP > CLOSE || KEEPALIVE_TIMEOUT) > > { > > if( this was our all neighbor connection) > > rejoin network > > > > ............ > > ............. > > } > > > > it this correct? > > Please see my message with timestamp "Sat 24 Mar 16:49". > The exactly logic is not important and depends on your > implementation. > -- > Bill Cheng // bill.cheng@usc.edu > > > > ----- Original Message ----- > From: william@bourbon.usc.edu > Date: Saturday, March 24, 2007 9:39 am > Subject: Re: check msg > To: cs551@merlot.usc.edu > > > Someone wrote: > > > > > We so not send check msg if a remote end closes its TCP > > > connection with out having sent NOTIFY. True? > > > > If you lose an *active* connection, you should flood a CHECK, > > no matter if you have received a NOTIFY or not. If the other > > side crashes, you should also flood a CHECK to see if you are > > in a partitioned SERVANT network. > > > > > What if it was the only neighbor of the node? What should > happen? > > > If I remember correctly, you delete your init_neighbor_list > > and rejoin the network. > > -- > > Bill Cheng // bill.cheng@usc.edu > > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 21:48:21 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4mLvX012176 for ; Sat, 24 Mar 2007 21:48:21 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P4iuaY002495 for ; Sat, 24 Mar 2007 21:44:56 -0700 Message-Id: <200703250444.l2P4iuaY002495@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: ctrl-c effect Date: Sat, 24 Mar 2007 21:44:56 -0700 From: william@bourbon.usc.edu Someone wrote: > what about check? Do we take the prompt away when check has been > sent? No. Since this node can be fine, so no need to take away the prompt. > Does pressing CTRL_C result in not waiting for check > response? If yes what should the node do next? Nothing or > re-join? JOIN is a bit different from CHECK. If you are doing a JOIN, you are not part of the network, so you can't really do anything else. If you are doing a CHECK, it's possible you can be doing a STATUS or something else. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 5:56 pm Subject: Re: ctrl-c effect To: cs551@merlot.usc.edu > Someone wrote: > > > Ctrl-C only has effect on status msg in this part, right? > > You should be able to press any time and your > program show not quit and you should get your command > prompt back. If you are doing a join and if you let the > user know that's what you are doing, you can ask the user > to press to signal end of waiting for join > responses and continue. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 17:58:48 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2P0wlIm026635 for ; Sat, 24 Mar 2007 17:58:47 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2P0tO4w000827 for ; Sat, 24 Mar 2007 17:55:24 -0700 Message-Id: <200703250055.l2P0tO4w000827@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: ctrl-c effect Date: Sat, 24 Mar 2007 17:55:24 -0700 From: william@bourbon.usc.edu Someone wrote: > Ctrl-C only has effect on status msg in this part, right? You should be able to press any time and your program show not quit and you should get your command prompt back. If you are doing a join and if you let the user know that's what you are doing, you can ask the user to press to signal end of waiting for join responses and continue. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:55:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONtIWx021647 for ; Sat, 24 Mar 2007 16:55:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONpsEO032705 for ; Sat, 24 Mar 2007 16:51:54 -0700 Message-Id: <200703242351.l2ONpsEO032705@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 16:51:54 -0700 From: william@bourbon.usc.edu Someone wrote: > if ( NOTIFY recvd on active connection || REMOTE SIDE TCP CLOSE || KEEPALIVE_TIMEOUT) > { > if( this was our all neighbor connection) > rejoin network > > ............ > ............. > } > > it this correct? Please see my message with timestamp "Sat 24 Mar 16:49". The exactly logic is not important and depends on your implementation. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 9:39 am Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > We so not send check msg if a remote end closes its TCP > > connection with out having sent NOTIFY. True? > > If you lose an *active* connection, you should flood a CHECK, > no matter if you have received a NOTIFY or not. If the other > side crashes, you should also flood a CHECK to see if you are > in a partitioned SERVANT network. > > > What if it was the only neighbor of the node? What should happen? > > If I remember correctly, you delete your init_neighbor_list > and rejoin the network. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:52:33 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONqXAM021477 for ; Sat, 24 Mar 2007 16:52:33 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONn9XI032662 for ; Sat, 24 Mar 2007 16:49:09 -0700 Message-Id: <200703242349.l2ONn9XI032662@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 16:49:09 -0700 From: william@bourbon.usc.edu Someone wrote: > does it mean that since the connection is no longer considered > active once NOTIFY is received on it, so we do not need to send > check? I'm not understanding what you are saying. Let's say that you are a regular node A and there is a connection to node B. If you lose this connection, node A must flood a CHECK. The timing is not very important. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 4:01 pm Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > do we flood a check if we get a NOTIFY on an active > > connection? > > Well, you should do it when you lose or close the connection. > If you are in the middle of sending data to this neighbor > which you just get a NOTIFY from, then I guess you need to > wait. So, it's probably easier to just do it after this > connection is removed. > -- > Bill Cheng // bill.cheng@usc.edu > > > > > ----- Original Message ----- > From: william@bourbon.usc.edu > Date: Saturday, March 24, 2007 9:39 am > Subject: Re: check msg > To: cs551@merlot.usc.edu > > > Someone wrote: > > > > > We so not send check msg if a remote end closes its TCP > > > connection with out having sent NOTIFY. True? > > > > If you lose an *active* connection, you should flood a CHECK, > > no matter if you have received a NOTIFY or not. If the other > > side crashes, you should also flood a CHECK to see if you are > > in a partitioned SERVANT network. > > > > > What if it was the only neighbor of the node? What should > happen? > > > If I remember correctly, you delete your init_neighbor_list > > and rejoin the network. > > -- > > Bill Cheng // bill.cheng@usc.edu > > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:13:35 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONDZxI018363 for ; Sat, 24 Mar 2007 16:13:35 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONAC52032306 for ; Sat, 24 Mar 2007 16:10:12 -0700 Message-Id: <200703242310.l2ONAC52032306@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: non-beacon node with insufficient neighbours Date: Sat, 24 Mar 2007 16:10:12 -0700 From: william@bourbon.usc.edu Someone wrote: > In case of non-beacon node joining the network for the first time > with no. of join responses less than initneighbours do the node > needs to restart and redo join process or it should die? Just quit and tell the user why you are quitting. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:12:27 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONCRlv018346 for ; Sat, 24 Mar 2007 16:12:27 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON93lb032244 for ; Sat, 24 Mar 2007 16:09:03 -0700 Message-Id: <200703242309.l2ON93lb032244@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 16:09:03 -0700 From: william@bourbon.usc.edu Someone wrote: > if case of the duplicate connections between Node A and Node B > (due to simultaneous starts) when one of the connection is closed > does one side flood a CHECK message when the other decision > making node terminates the duplicate connection? Since the reason for CHECK is to see if there is a network partition, there is no need to send a CHECK in this case since you really haven't lost a connection with a neighbor. You were just killing an extra connection with a neighbor. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 9:39 am Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > We so not send check msg if a remote end closes its TCP > > connection with out having sent NOTIFY. True? > > If you lose an *active* connection, you should flood a CHECK, > no matter if you have received a NOTIFY or not. If the other > side crashes, you should also flood a CHECK to see if you are > in a partitioned SERVANT network. > > > What if it was the only neighbor of the node? What should happen? > > If I remember correctly, you delete your init_neighbor_list > and rejoin the network. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:10:34 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ONAYeR018179 for ; Sat, 24 Mar 2007 16:10:34 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON7Bkq032209 for ; Sat, 24 Mar 2007 16:07:11 -0700 Message-Id: <200703242307.l2ON7Bkq032209@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 16:07:11 -0700 From: william@bourbon.usc.edu Someone wrote: > does a beacon node initiate check msgs? Since the point of a CHECK message is to check for network partition, and beacon nodes always form a full mesh, you never need to send a CHECK from a beacon node. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 9:39 am Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > We so not send check msg if a remote end closes its TCP > > connection with out having sent NOTIFY. True? > > If you lose an *active* connection, you should flood a CHECK, > no matter if you have received a NOTIFY or not. If the other > side crashes, you should also flood a CHECK to see if you are > in a partitioned SERVANT network. > > > What if it was the only neighbor of the node? What should happen? > > If I remember correctly, you delete your init_neighbor_list > and rejoin the network. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:09:16 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON9GjC018007 for ; Sat, 24 Mar 2007 16:09:16 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON5qPK032177 for ; Sat, 24 Mar 2007 16:05:52 -0700 Message-Id: <200703242305.l2ON5qPK032177@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: status neighbors Date: Sat, 24 Mar 2007 16:05:52 -0700 From: william@bourbon.usc.edu Someone wrote: > is the name of an external file (relative to the > current working directory if it does not begin with a / > character). > > So if the extfile is passed as directoryname/namfile.out > then are we supposed to create the directory called directoryname > if it doesnt exist? Or you can just print out an error message and ask the user to try another filename. This way, your code works no matter if the directory specified is in a good place or a bad place (such as /usr). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 16:04:00 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON40mZ017536 for ; Sat, 24 Mar 2007 16:04:00 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2ON0alk032089 for ; Sat, 24 Mar 2007 16:00:36 -0700 Message-Id: <200703242300.l2ON0alk032089@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 16:00:36 -0700 From: william@bourbon.usc.edu Someone wrote: > do we flood a check if we get a NOTIFY on an active > connection? Well, you should do it when you lose or close the connection. If you are in the middle of sending data to this neighbor which you just get a NOTIFY from, then I guess you need to wait. So, it's probably easier to just do it after this connection is removed. -- Bill Cheng // bill.cheng@usc.edu ----- Original Message ----- From: william@bourbon.usc.edu Date: Saturday, March 24, 2007 9:39 am Subject: Re: check msg To: cs551@merlot.usc.edu > Someone wrote: > > > We so not send check msg if a remote end closes its TCP > > connection with out having sent NOTIFY. True? > > If you lose an *active* connection, you should flood a CHECK, > no matter if you have received a NOTIFY or not. If the other > side crashes, you should also flood a CHECK to see if you are > in a partitioned SERVANT network. > > > What if it was the only neighbor of the node? What should happen? > > If I remember correctly, you delete your init_neighbor_list > and rejoin the network. > -- > Bill Cheng // bill.cheng@usc.edu > Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 15:59:30 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMxUM9017200 for ; Sat, 24 Mar 2007 15:59:30 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMu6wC032019 for ; Sat, 24 Mar 2007 15:56:06 -0700 Message-Id: <200703242256.l2OMu6wC032019@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj Date: Sat, 24 Mar 2007 15:56:06 -0700 From: william@bourbon.usc.edu Someone wrote: > Should we use gettimeofday() in the time field in logging. Or anything that can give you millisecond resolution since the logging format requires 3 digits after the decimal point. > Should we just call gettimeofday() while logging or should we call it > once at the start of the node and then every log entry is > relative to that recorded at the start. This is not specified in the spec. So, you can do it any way you like, as long as it's consistent. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 15:52:09 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMq9Uh016716 for ; Sat, 24 Mar 2007 15:52:09 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMmjxx031874 for ; Sat, 24 Mar 2007 15:48:45 -0700 Message-Id: <200703242248.l2OMmjxx031874@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Time field in the ServantLog Date: Sat, 24 Mar 2007 15:48:45 -0700 From: william@bourbon.usc.edu Someone wrote: > Does the time field in the Servant Logs indicate the current time > or the time from the moment the node has started.. The spec doesn't say which. So, whatever you'd like. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 15:45:39 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMjd7e016073 for ; Sat, 24 Mar 2007 15:45:39 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2OMgFeU031776 for ; Sat, 24 Mar 2007 15:42:15 -0700 Message-Id: <200703242242.l2OMgFeU031776@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Updating the ServantLogs Date: Sat, 24 Mar 2007 15:42:15 -0700 From: william@bourbon.usc.edu Someone wrote: > When a node shutdown and restart do we need to remove the > servantlogs or update the existing old servantlog. You don't need to delete it. (If you delete it when you do a real restart, i.e., not a soft-restart, I think it's fine too.) -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 09:41:52 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2OGfqDZ019915 for ; Sat, 24 Mar 2007 09:41:52 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2OGcTfd028673 for ; Sat, 24 Mar 2007 09:38:29 -0700 Message-Id: <200703241638.l2OGcTfd028673@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: check msg Date: Sat, 24 Mar 2007 09:38:29 -0700 From: william@bourbon.usc.edu Someone wrote: > We so not send check msg if a remote end closes its TCP > connection with out having sent NOTIFY. True? If you lose an *active* connection, you should flood a CHECK, no matter if you have received a NOTIFY or not. If the other side crashes, you should also flood a CHECK to see if you are in a partitioned SERVANT network. > What if it was the only neighbor of the node? What should happen? If I remember correctly, you delete your init_neighbor_list and rejoin the network. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Sat Mar 24 09:37:41 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2OGbfri019585 for ; Sat, 24 Mar 2007 09:37:41 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2OGYJfL028609 for ; Sat, 24 Mar 2007 09:34:19 -0700 Message-Id: <200703241634.l2OGYJfL028609@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: What is the default value for Date: Sat, 24 Mar 2007 09:34:19 -0700 From: william@bourbon.usc.edu Someone wrote: > You said, this "extfile" is required parameter. Though, TTL is not. Right? > Actually, I am not sure about it. > So, please clarify this. > > If TTL is not mandatory, how do we know that the 3rd parameter is a > filename. As a filename, can also be only digits! > I hope you understand what I mean. TTL is also required. If something is optional, it would be surrounded by [] in the spec, according to UNIX conventions. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Friday, March 23, 2007 10:15 PM To: cs551@merlot.usc.edu Subject: Re: What is the default value for Someone wrote: > Another quick question. > > What is the default value for on command shell? > Is it displayed on screen? is mandatory, so there is no default value. Please see the grading guidelines for examples. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 22:32:13 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5WDRj031849 for ; Fri, 23 Mar 2007 22:32:13 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5Sq09022439 for ; Fri, 23 Mar 2007 22:28:52 -0700 Message-Id: <200703240528.l2O5Sq09022439@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: fgets Date: Fri, 23 Mar 2007 22:28:52 -0700 From: william@bourbon.usc.edu Someone wrote: > fgets() does not seem to respond to signals. Do you > recommend an alternative? You can write your own fgets() using read(). -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 22:28:37 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5Sb8H031521 for ; Fri, 23 Mar 2007 22:28:37 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5PGWd022324 for ; Fri, 23 Mar 2007 22:25:16 -0700 Message-Id: <200703240525.l2O5PGWd022324@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: NoCheck and reset Date: Fri, 23 Mar 2007 22:25:16 -0700 From: william@bourbon.usc.edu Someone wrote: > Do we need to implement NoCheck and reset command line > options in final project part 1? > I am asking because in grading guidlines I couldnt find any > test-case using them. You don't need to implement NoCheck for part (1). How are you cleaning up your node so far if you don't have -reset implemented? Anyway, I've updated the grading guidelines to say to take up to 5 points off if -reset is not implemented. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 22:19:55 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5JtEM030664 for ; Fri, 23 Mar 2007 22:19:55 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5GYej022174 for ; Fri, 23 Mar 2007 22:16:34 -0700 Message-Id: <200703240516.l2O5GYej022174@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Forwarding msgs Date: Fri, 23 Mar 2007 22:16:34 -0700 From: william@bourbon.usc.edu Someone wrote: > This is actually a fundamental doubt. Any msg that mayb potentially > forwarded anywhere in the network, has to be added to the forwarding table > of the node before it is transmitted. Specifically, will this message be > added to the table at the origin node? Of course! Otherwise, how do you prevent looping? -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 22:18:44 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5IijJ030648 for ; Fri, 23 Mar 2007 22:18:44 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5FNPJ022146 for ; Fri, 23 Mar 2007 22:15:23 -0700 Message-Id: <200703240515.l2O5FNPJ022146@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: What is the default value for Date: Fri, 23 Mar 2007 22:15:23 -0700 From: william@bourbon.usc.edu Someone wrote: > Another quick question. > > What is the default value for on command shell? > Is it displayed on screen? is mandatory, so there is no default value. Please see the grading guidelines for examples. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 22:04:02 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O5416N029403 for ; Fri, 23 Mar 2007 22:04:01 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O50eTN021981 for ; Fri, 23 Mar 2007 22:00:40 -0700 Message-Id: <200703240500.l2O50eTN021981@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: A value of zero means not to flood the status message. Date: Fri, 23 Mar 2007 22:00:40 -0700 From: william@bourbon.usc.edu Someone wrote: > What do you mean by this? > A value of zero means not to flood the status message. If you enter "status neighbors 0 status.out", you should not flood a status request to your neighbors. So, the only thing that goes into "status.out" is the neighbor information of the node you are on. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 21:58:51 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2O4wp3t028916 for ; Fri, 23 Mar 2007 21:58:51 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2O4tU0T021870 for ; Fri, 23 Mar 2007 21:55:30 -0700 Message-Id: <200703240455.l2O4tU0T021870@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: CNTR C Date: Fri, 23 Mar 2007 21:55:30 -0700 From: william@bourbon.usc.edu Someone wrote: > we were having some difficulty with handling CNTRL C. Basically we have > gotten to a point where if a user presses CNTL C during the execution of a > user command, we get a promp however when a user presses CNTRL C while on > the prompt the program exits. I think all you have to do is to call signal() again to get the signal handler back what you were using. I'm not 100% sure about this, but I think it's something like that. > Are we not handling SIGINT properly or is it supposed to behave this way? > Is this acceptable from a design point, I mean can we give this as a design > decision? is not suppose to kill your node. For commands that needs to wait, is a way to say "don't wait any more, let's see what you've got". Also, it's a way to "get the command prompt back". I see a parallel between the commandline interface of a node and the UNIX shell (csh/bash). In csh/bash, if you run some command and it's not returning, you can press to get the UNIX prompt back (and kill whatever you were doing). For our node, when you press , you should also get the commandline prompt back, but you don't kill whatever you were doing, you force it to finish. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 16:38:56 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2NNcuBK003653 for ; Fri, 23 Mar 2007 16:38:56 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2NNZaet015332 for ; Fri, 23 Mar 2007 16:35:36 -0700 Message-Id: <200703232335.l2NNZaet015332@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: How to typecast a function? Date: Fri, 23 Mar 2007 16:35:36 -0700 From: william@bourbon.usc.edu Someone wrote: > I am not able to get rid of this type of warning: (for pthread) > src/main.c:126: warning: passing arg 3 of `pthread_create' from incompatible > pointer type > > Where argument 3 is the thread function I am assigning for pthread. > > I think a typecast will be able to solve the problem, but I don't know how. > > Function prototype: > int pthread_create(pthread_t * thread, pthread_attr_t * attr, void > *(*start_routine)(void *), void * arg); If you declare your start routine correctly, you should not have the problem. For example: void *MyStartRoutine(void *ptr) { ... } If you have to use something different, say: int MyOtherStartRoutine(int) { ... } Then you can do something like the following: typedef void *(StartRoutineType)(void*); pthread_create(..., (StartRoutineType*)MyOtherStartRoutine); But, the compiler is telling you the right thing here! You *should* make your thread start routine conform to the way pthread expects you to! -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Friday, March 23, 2007 4:19 PM To: cs551@merlot.usc.edu Subject: Re: How to typecast a function? Someone wrote: > I am facing a problem with function typecasting. Pthread generates warning > for different types of function in parameter. I wanted to typecast it, but > don't know how can I do that. > > Will appreciate if you could help. I'm not sure what you want to typecast. Let's take qsort() as an example. The man pages say: void qsort(void *base, size_t nmemb, size_t size, int(*compar)(const void *, const void *)); So, you can write the following function: static int my_compare(const void *p2, const void *p2) { ... } Then you can call: qsort(..., my_compare); So, there is no need to typecast if you start with the right function. This is probably not what you are looking for. Please let me know what kind of typecasting you need. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 16:22:01 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2NNM1QN002216 for ; Fri, 23 Mar 2007 16:22:01 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2NNIfCJ021162 for ; Fri, 23 Mar 2007 16:18:41 -0700 Message-Id: <200703232318.l2NNIfCJ021162@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: How to typecast a function? Date: Fri, 23 Mar 2007 16:18:41 -0700 From: william@bourbon.usc.edu Someone wrote: > I am facing a problem with function typecasting. Pthread generates warning > for different types of function in parameter. I wanted to typecast it, but > don't know how can I do that. > > Will appreciate if you could help. I'm not sure what you want to typecast. Let's take qsort() as an example. The man pages say: void qsort(void *base, size_t nmemb, size_t size, int(*compar)(const void *, const void *)); So, you can write the following function: static int my_compare(const void *p2, const void *p2) { ... } Then you can call: qsort(..., my_compare); So, there is no need to typecast if you start with the right function. This is probably not what you are looking for. Please let me know what kind of typecasting you need. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 13:47:12 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2NKlCRL022585 for ; Fri, 23 Mar 2007 13:47:12 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2NKhqEB016012 for ; Fri, 23 Mar 2007 13:43:52 -0700 Message-Id: <200703232043.l2NKhqEB016012@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Proj Date: Fri, 23 Mar 2007 13:43:52 -0700 From: william@bourbon.usc.edu Someone wrote: > I was performing the cleanup operation before shutdown. I was > trying something and i forgot to close the sockets and i ran the > code a couple of times and now if i run the code it says "cannot > open socket: too many open files" returned by function socket(). > What should i do?? It is an error no EMFILE when i google it up > so what do i d?? I'm not sure what are you asking about. Is this still the same process? Or this is about after killing your node you cannot restart? If you have a port you cannot seem to reuse (for one reason or another), you can just use another port! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Fri Mar 23 13:29:29 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2NKTTgl021173 for ; Fri, 23 Mar 2007 13:29:29 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2NKQAY7015765 for ; Fri, 23 Mar 2007 13:26:10 -0700 Message-Id: <200703232026.l2NKQAY7015765@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: If JoinTimeout>KeepAliveTimeout Date: Fri, 23 Mar 2007 13:26:10 -0700 From: william@bourbon.usc.edu Someone wrote: > What is normal behavior if the JoinTimeout > KeepAliveTimeout value. The > Beacon may close the connection as the child is no longer respnsive and > waiting for JNRS. > > One way to solve is keep sending KPAV at regular intervals from the node to > beacon, making it sure that it can never close the connection then. You are talking about the case where a non-beacon node is trying to join the network. Since this node is not part of the network yet, the connection it uses to talk to the beacon node is a "special connection". You don't need to send KeepAlive messages on a "special connection". -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 23:17:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N6HItK018082 for ; Thu, 22 Mar 2007 23:17:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N6E0ds006447 for ; Thu, 22 Mar 2007 23:14:00 -0700 Message-Id: <200703230614.l2N6E0ds006447@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Date: Thu, 22 Mar 2007 23:14:00 -0700 From: william@bourbon.usc.edu Someone wrote: > I am sorry, but I was asking you about logging hello messages in the log > file. We do log those right? Yes. Of course (see grading guidelines). -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Thursday, March 22, 2007 11:05 PM To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Someone wrote: > Oh Ok! But there is one situation where I am not able to get that behavior. > > Actually, my code is not wating for incoming hello msg to reply with a hello > msg. so, in my case the sending log entry (s ...) will not know the > destination's host/port (as it wasn't waiting for incoming hello). Should I > consider this behavior as wrong? If you have not exchanged HELLO messages on a connection, it should not be considered an *active* connection. You should not forward any messages on an inactive connection. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Thursday, March 22, 2007 10:49 PM To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Someone wrote: > Another quick question. > > What do we print in the log file, if the from / to field is not known > (host/port is not known). Is it okay to have (null)_0 written in the log for > those enteries? You are suppose to know! This is because if a connection is active, then it must mean that you have exchanged HELLO messages on this connection. Therefore, you must have information about host/port. You just need to keep it so you can print it. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 23:07:51 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N67pLc017290 for ; Thu, 22 Mar 2007 23:07:51 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N64YqW006191 for ; Thu, 22 Mar 2007 23:04:34 -0700 Message-Id: <200703230604.l2N64YqW006191@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Date: Thu, 22 Mar 2007 23:04:34 -0700 From: william@bourbon.usc.edu Someone wrote: > Oh Ok! But there is one situation where I am not able to get that behavior. > > Actually, my code is not wating for incoming hello msg to reply with a hello > msg. so, in my case the sending log entry (s ...) will not know the > destination's host/port (as it wasn't waiting for incoming hello). Should I > consider this behavior as wrong? If you have not exchanged HELLO messages on a connection, it should not be considered an *active* connection. You should not forward any messages on an inactive connection. -- Bill Cheng // bill.cheng@usc.edu -----Original Message----- From: william@bourbon.usc.edu [mailto:william@bourbon.usc.edu] Sent: Thursday, March 22, 2007 10:49 PM To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Someone wrote: > Another quick question. > > What do we print in the log file, if the from / to field is not known > (host/port is not known). Is it okay to have (null)_0 written in the log for > those enteries? You are suppose to know! This is because if a connection is active, then it must mean that you have exchanged HELLO messages on this connection. Therefore, you must have information about host/port. You just need to keep it so you can print it. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 22:52:31 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5qVrK016120 for ; Thu, 22 Mar 2007 22:52:31 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5nEYm005923 for ; Thu, 22 Mar 2007 22:49:14 -0700 Message-Id: <200703230549.l2N5nEYm005923@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: From/to filed in log file Date: Thu, 22 Mar 2007 22:49:14 -0700 From: william@bourbon.usc.edu Someone wrote: > Another quick question. > > What do we print in the log file, if the from / to field is not known > (host/port is not known). Is it okay to have (null)_0 written in the log for > those enteries? You are suppose to know! This is because if a connection is active, then it must mean that you have exchanged HELLO messages on this connection. Therefore, you must have information about host/port. You just need to keep it so you can print it. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 22:32:58 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5WvRR014543 for ; Thu, 22 Mar 2007 22:32:58 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5Temq005534 for ; Thu, 22 Mar 2007 22:29:40 -0700 Message-Id: <200703230529.l2N5Temq005534@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: status neighbor response Date: Thu, 22 Mar 2007 22:29:40 -0700 From: william@bourbon.usc.edu Someone wrote: > When a node receives a status neighbor message does it > immediately respond with a response stating its port and > hostname? (and then forward the msg if TTL is not zero after > decrementing it) Yes, and report the ports and hostnames of all its neighbors. > In that case each response to status will have exactly one > port:hostname pair. I'm not sure why you said "exactly one". For a response that's sent by node X, the response contain the ports and hostnames of all the active neighbors of node X. > How does the node which originally sent out > the status neighbor msg find out what the path to the responding > node is? It doesn't need the path since the graph can be constructed if you have all the links (and each link is described by the names of its two endpoints). > In this case it will not be able to create the nam > format external file correctly, right? I think it could. If you just use the simple algorithm I mentioned in my message with timestamp "Thu 22 Mar 21:51", I think it would work. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 22:23:37 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5NbnS013762 for ; Thu, 22 Mar 2007 22:23:37 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N5KKYG005431 for ; Thu, 22 Mar 2007 22:20:20 -0700 Message-Id: <200703230520.l2N5KKYG005431@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: What is the maximum length of a domain name? Date: Thu, 22 Mar 2007 22:20:20 -0700 From: william@bourbon.usc.edu Someone wrote: > I have a quick question. > What is the maximum length of a host/domain name for our project? Since we will only test on nunki, using 256 bytes should be safe! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 21:57:48 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4vmiH011559 for ; Thu, 22 Mar 2007 21:57:48 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4sVua004981 for ; Thu, 22 Mar 2007 21:54:31 -0700 Message-Id: <200703230454.l2N4sVua004981@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: problem in fgets Date: Thu, 22 Mar 2007 21:54:31 -0700 From: william@bourbon.usc.edu Someone wrote: > we are trying to implement the command prompt using fgets by > calling it repeatedly from a seperate thread. But it looks like > that it doesnt wait for the user input and infinitely prints the > command prompt. Kindly suggest the solution to this problem. Is > this a problem related with Pthread implementation.. Hmm... This sounds familiar. I think I've heard it from previous semesters. But they were all fixed; although I don't recalling anyone telling me what the fix was. If fgets() doesn't work but fread() one byte at a time works, you can write your own fgets() using fread() one byte at a time. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 21:54:26 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4sQc4011383 for ; Thu, 22 Mar 2007 21:54:26 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4p9Ym004911 for ; Thu, 22 Mar 2007 21:51:09 -0700 Message-Id: <200703230451.l2N4p9Ym004911@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Regarding Test cases Date: Thu, 22 Mar 2007 21:51:09 -0700 From: william@bourbon.usc.edu Someone wrote: > The first test case states > > in one window, do: > ~csci551b/bin/resetf1 > /bin/rm n??.ini ??.out > cp ~csci551b/public/final1/b1-n??.ini . > ./sv_node b1-n00.ini > (wait 2 seconds, at the commandline prompt, type > "status neighbors 1 00.out", look at "00.out"; > this file should have no links) > > here i do not get the file 00.out . > > I have a wait time(i mean status time out) of 1.5*message Life > Time for status command. I mean to get all replies. Then i write > to a file. > > Default message life time is 30. so to generate the file it will > take 45 secs for me. > In this case since auto shutdown is 10 secs the node shuts down > before the file is written > > But the test case specifies 00.out should have no links. But my > program doesn't generate the file itself. > > Please suggest me what should i do I think I have mentioned this in class. You should not wait for all responses to come in before you start doing the output. (In general, you should do this for all commands when appropriate.) For the status response, you need to keep a list of nodes that you have seen for this command. If you get a response mentioning a node that's not on your list, you should output this node immediately and add it to your list. Then whenever you see a link in a response, you should output it immediately. Of course, you need to do something special for the node that initiated the command since it does not send a message to itself. This should be very straightfoward. You just output this node and all it's neighbors, then output all the links to all its neighbors. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 21:44:57 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4ivQU010595 for ; Thu, 22 Mar 2007 21:44:57 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4feSl004781 for ; Thu, 22 Mar 2007 21:41:40 -0700 Message-Id: <200703230441.l2N4feSl004781@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Errors in grading guidelines - Number of HLLO messages Date: Thu, 22 Mar 2007 21:41:40 -0700 From: william@bourbon.usc.edu Someone wrote: > There are some errors (according to me) in the grading guidelines. We (me > and project partner) spent a lot of time discussing the errors and concluded > that the nodes may not output the log file as mentioned in the guidelines. > > I am referring to the number of HLLO messages for first 4 test cases (ie. > Beacons only networks) except the 1 beacon network. > > For Eg. > > The spec for 4 beacon network says that: > > in 5th window, do: > more ~/tmp/final1/n00/servant.log > more ~/tmp/final1/n01/servant.log > more ~/tmp/final1/n02/servant.log > more ~/tmp/final1/n03/servant.log > (all these files should have things like HLLO, KPAV, > NTFY, > STRQ, and STRS; there should be 6 HLLOs, 3 on > sending > and 3 on receiving) > > You have mentioned that there will should be 6 HLLOs. But, there can be a > possiblity that there are more of them. The correct thing may be - "there > should be 6 or more HLLOs". As the retry interval in your b4-xx.ini (4 > beacon ini files) is 5 seconds, there is high probability that the new node > started is at the 5th second. This may result in two connections initially, > and later one of the connections is dropped by the nodes. But, this has > eventually increased the number of HLLO messages. I think you are correct. Some nodes may be retrying at exactly the same time as you start a node. So, it's possible to end up with simultaneous connection and get more HELLO messages. One way to fix this is to change the Retry timeout to be a much larger value. Another way is to do what you have suggested. In order not to change everywhere, I've added an item (7) to the general guidelines: 7) Due to unpredictable timing, if the grading guidelines says "there should be N number of HLLO messages" in the log file, please treat this as saying "there should be N or more HLLO messages". > I am sorry that I am pointing out an error with your spec. If there is > something wrong with my understanding, will appreciate if you could > clarify. There is no need to apologize. I strongly encourage all of you to find bugs in the grading guidelines! Thanks for catching the bugs! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 21:25:59 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4PxWj008986 for ; Thu, 22 Mar 2007 21:25:59 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2N4Mgd6004516 for ; Thu, 22 Mar 2007 21:22:42 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l2N4MgZh004515 for cs551@merlot; Thu, 22 Mar 2007 21:22:42 -0700 Date: Thu, 22 Mar 2007 21:22:42 -0700 From: william@bourbon.usc.edu Message-Id: <200703230422.l2N4MgZh004515@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Project partner commitment, time is running out! Hi, If you are doing final project part (1) with a partner, both of you should send me an e-mail letting me know about it *now*. In your e-mail, please copy it to your partner. You don't need to copy the e-mail to the TA or the grader. Once the submission deadline has past and I started to run MOSS to check for plagiarism. If I see that your code is too similar to another student's code, then it's *too late* to say, "oh, I forgot to mention, he/she is my partner". If you do not plan to submit a written commitment, please include the following two paragraphs in your e-mail: I agree to receive the same score as my partner no matter what happens in the remainder of the Spring 2007 semester. This includes, but not limited to, the case where my partner does not make any contribution to the submitted project, drops the course, or withdraws from the semester. I was not forced in any way to do a group project. I fully understand that all projects, done individually or by a group of two students, will be graded in an identical manner, even if the group project is done completely by one student. If you plan to submit a written commitment, please let me know in your e-mail. Thanks! -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 11:06:40 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2MI6eJk025089 for ; Thu, 22 Mar 2007 11:06:40 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2MI2G0W030161 for ; Thu, 22 Mar 2007 11:02:16 -0700 Received: (from william@localhost) by bourbon.usc.edu (8.13.5/8.13.5/Submit) id l2MI2GTn030160 for cs551@merlot; Thu, 22 Mar 2007 11:02:16 -0700 Date: Thu, 22 Mar 2007 11:02:16 -0700 From: william@bourbon.usc.edu Message-Id: <200703221802.l2MI2GTn030160@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: We will talk about final project part (2) today Hi, Sorry about the short notice. But I just realize that today is probably a good day to talk about part (2) since part (1) is due this Saturday. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 10:37:18 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHbIXX022716 for ; Thu, 22 Mar 2007 10:37:18 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHWtMk029739 for ; Thu, 22 Mar 2007 10:32:55 -0700 Message-Id: <200703221732.l2MHWtMk029739@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Joining fail and retry Date: Thu, 22 Mar 2007 10:32:55 -0700 From: william@bourbon.usc.edu Someone wrote: > so if a node has 2 beacon nodes in the ini file and it attempts > the join through the first one and cant get enough join responses > (so the TCP connect with beacon goes through fine) it should not > try the next beacon node but instead exit? It doesn't need to. The assumption is that the beacon nodes form a full mesh. Network partition is not possible given our protocol (other than transients). If you want to avoid the transients, I guess you can go down the list. I think it should not break anything in the grading guidelines. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 10:29:05 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHT5iI022066 for ; Thu, 22 Mar 2007 10:29:05 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHOgHD029525 for ; Thu, 22 Mar 2007 10:24:42 -0700 Message-Id: <200703221724.l2MHOgHD029525@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: Init:Location conversion Date: Thu, 22 Mar 2007 10:24:42 -0700 From: william@bourbon.usc.edu Someone wrote: > I am facing a problem after converting the string "3134382376" as int. I > tried using atoi/atol and even strtol but none of them work. They all are > returing 2147483647. > > I am not understand this problem of typecast. Will appreciate if you could > help. If you know that the string you are reading represents a number and that after you've read it, it fits into an integer (32-bits), then it's like you are trying to read this string representation of an interger from a file. So, you can use sscanf(). For example: unsigned int integer_value=0; sscanf("3134382376", "%d", &integer_value); printf("%u\n", integer_value); There is no "typecasting" here. -- Bill Cheng // bill.cheng@usc.edu Return-Path: william@bourbon.usc.edu Delivery-Date: Thu Mar 22 10:18:24 2007 X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on merlot.usc.edu X-Spam-Level: X-Spam-Status: No, score=-3.3 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00, NO_REAL_NAME autolearn=ham version=3.1.3 Received: from bourbon.usc.edu (bourbon.usc.edu [128.125.9.75]) by merlot.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHIOnW021243 for ; Thu, 22 Mar 2007 10:18:24 -0700 Received: from bourbon.usc.edu (localhost.localdomain [127.0.0.1]) by bourbon.usc.edu (8.13.5/8.13.5) with ESMTP id l2MHE0WC029298 for ; Thu, 22 Mar 2007 10:14:00 -0700 Message-Id: <200703221714.l2MHE0WC029298@bourbon.usc.edu> To: cs551@merlot.usc.edu Subject: Re: about logging Date: Thu, 22 Mar 2007 10:14:00 -0700 From: william@bourbon.usc.edu Someone wrote: > I'm currently coding logging part of the project. > For the logging part, the spec requires: > > "The