From comm at nrao.edu Tue May 25 13:30:03 2010 From: comm at nrao.edu (Communications Coordination Group) Date: Tue, 25 May 2010 13:30:03 -0400 Subject: [comm] Communications meeting Wednesday Message-ID: <201005251730.o4PHU3SX017017@polaris.cv.nrao.edu> There will be a communication meeting this Wednesday at 17:30UT/13:30ET/11:30MT. Locations: CV-conf (CV-209), GB-conf2 (GB-241), SO-conf (AOC-317), Santiago From grunion at nrao.edu Wed May 26 12:47:36 2010 From: grunion at nrao.edu (Gene Runion) Date: Wed, 26 May 2010 12:47:36 -0400 Subject: [comm] Reminder Message-ID: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> https://staff.nrao.edu/wiki/bin/view/Ccs/MonthlyCommunicationMeetings There will be a communication meeting this Wednesday at 17:30UT/13:30ET/11:30MT. Locations: CV-conf (CV-209), GB-conf2 (GB-241), SO-conf (AOC-317) Gene Runion 434-296-0327 From jrobnett at nrao.edu Wed May 26 13:05:08 2010 From: jrobnett at nrao.edu (James Robnett) Date: Wed, 26 May 2010 11:05:08 -0600 Subject: [comm] Reminder In-Reply-To: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> References: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> Message-ID: <4BFD54C4.3020506@aoc.nrao.edu> I may be a bit late. We're swapping some stuff on the correlator and this is our window. James Gene Runion wrote: > https://staff.nrao.edu/wiki/bin/view/Ccs/MonthlyCommunicationMeetings > > There will be a communication meeting this Wednesday at 17:30UT/13:30ET/11:30MT. > > Locations: CV-conf (CV-209), GB-conf2 (GB-241), SO-conf (AOC-317) > > Gene Runion > 434-296-0327 > > > > _______________________________________________ > Comm mailing list > Comm at listmgr.cv.nrao.edu > http://listmgr.cv.nrao.edu/mailman/listinfo/comm From jrobnett at nrao.edu Wed May 26 14:18:23 2010 From: jrobnett at nrao.edu (James Robnett) Date: Wed, 26 May 2010 12:18:23 -0600 Subject: [comm] Reminder In-Reply-To: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> References: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> Message-ID: <4BFD65EF.7030604@aoc.nrao.edu> Just did the network tests with NMT. Off hand it looks like there's a bad fiber somewhere between us since the tests ruled out the transceiver. We're getting no where near the throughput we should be getting over that 1gbit link to NMT and UNM. This *doesn't* explain the poor performance to CV, in the AOC to NMT/UNM case we're missing many tens of percent bandwidth. In the AOC to CV case we're missing 99.9% of the bandwidth. James Gene Runion wrote: > https://staff.nrao.edu/wiki/bin/view/Ccs/MonthlyCommunicationMeetings > > There will be a communication meeting this Wednesday at 17:30UT/13:30ET/11:30MT. > > Locations: CV-conf (CV-209), GB-conf2 (GB-241), SO-conf (AOC-317) > > Gene Runion > 434-296-0327 > > > > _______________________________________________ > Comm mailing list > Comm at listmgr.cv.nrao.edu > http://listmgr.cv.nrao.edu/mailman/listinfo/comm From dhalstea at nrao.edu Thu May 27 19:03:27 2010 From: dhalstea at nrao.edu (David Halstead) Date: Thu, 27 May 2010 19:03:27 -0400 Subject: [comm] Throughput on a lossy link In-Reply-To: <4BFD65EF.7030604@aoc.nrao.edu> References: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> <4BFD65EF.7030604@aoc.nrao.edu> Message-ID: <4BFEFA3F.5040207@nrao.edu> James, Gene and I were mulling this over and remembered reading about the bandwidth wall you hit with higher latency links with even minor packet loss. Baseline: To achieve a gigabit/sec throughput on a 1500MTU link with 40ms RTT you need a packet loss of less than 0.00001%!! A loss of 0.1% gives a max throughput of ~9Mbps for a single TCP stream, even for a 10Gig link of the same latency. For more info see: http://sd.wareonearth.com/~phil/issues.html and Not that this changes the plan to fix the local link to NMT, but that problem MAY also explain the 99.9% bandwidth loss to CV due to the 1/(rtt*sqrt(loss)) term. David James Robnett wrote: > Just did the network tests with NMT. Off hand it looks like > there's a bad fiber somewhere between us since the tests ruled out > the transceiver. > > We're getting no where near the throughput we should be getting over > that 1gbit link to NMT and UNM. > > This *doesn't* explain the poor performance to CV, in the AOC to > NMT/UNM case we're missing many tens of percent bandwidth. In the > AOC to CV case we're missing 99.9% of the bandwidth. > > James > > Gene Runion wrote: > >> https://staff.nrao.edu/wiki/bin/view/Ccs/MonthlyCommunicationMeetings >> >> There will be a communication meeting this Wednesday at 17:30UT/13:30ET/11:30MT. >> >> Locations: CV-conf (CV-209), GB-conf2 (GB-241), SO-conf (AOC-317) >> >> Gene Runion >> 434-296-0327 >> >> >> >> _______________________________________________ >> Comm mailing list >> Comm at listmgr.cv.nrao.edu >> http://listmgr.cv.nrao.edu/mailman/listinfo/comm >> > _______________________________________________ > Comm mailing list > Comm at listmgr.cv.nrao.edu > http://listmgr.cv.nrao.edu/mailman/listinfo/comm > From jrobnett at nrao.edu Fri May 28 10:02:26 2010 From: jrobnett at nrao.edu (James Robnett) Date: Fri, 28 May 2010 08:02:26 -0600 Subject: [comm] Throughput on a lossy link In-Reply-To: <4BFEFA3F.5040207@nrao.edu> References: <23A6C4404FF54442BB9703F53D255350012F5D8956@cv-er-xsb> <4BFD65EF.7030604@aoc.nrao.edu> <4BFEFA3F.5040207@nrao.edu> Message-ID: <4BFFCCF2.3080908@aoc.nrao.edu> I don't think his math is quite correct. It assumes 1 to 1 ack correspondence and doesn't take into account larger windows sizes and whether selective acks are enabled (tcp_sack). Certainly packet loss is murder but I don't think it explains the entire problem. We're clearly suffering packet loss here but once that's solved I think we'll still have other issues. Our loss doesn't explain the slow rate from CV to UNM for instance. We're doing more tests now but I think we're just tackling the first of many problems. james