From owner-freebsd-chat Wed Feb 21 14:53:38 1996 Return-Path: owner-chat Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id OAA06656 for chat-outgoing; Wed, 21 Feb 1996 14:53:38 -0800 (PST) Received: from Root.COM (implode.Root.COM [198.145.90.17]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id OAA06651 for ; Wed, 21 Feb 1996 14:53:36 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by Root.COM (8.6.12/8.6.5) with SMTP id OAA01582; Wed, 21 Feb 1996 14:53:18 -0800 Message-Id: <199602212253.OAA01582@Root.COM> X-Authentication-Warning: implode.Root.COM: Host localhost didn't use HELO protocol To: Charles Henrich cc: jkh@time.cdrom.com, freebsd-chat@freebsd.org Subject: Re: Woah! cdrom.com! In-reply-to: Your message of "Wed, 21 Feb 1996 14:34:16 EST." <199602211934.OAA01317@crh.cl.msu.edu> From: David Greenman Reply-To: davidg@Root.COM Date: Wed, 21 Feb 1996 14:53:18 -0800 Sender: owner-chat@freebsd.org Precedence: bulk >>In any case, it actually *IS* at 100Mbits now. CRL evidently >>purchased another type of 100Mbit card for their Cisco and now we're >>frequently peaking over 10Mbits, albeit rarely. > >>It seems that 600 users eat up around 10Mbits and not much more, so >>it's more a matter of "however fast you can go, we can go" now.. :-) > >There has to be a bottle neck somewhere. I have a T3 that is peaking out at >8MBit/sec, and I have a full ethernet to that T3. > >And my retrieval performance is *dismal* > >Retrieving boot.flp from cdrom.com is ETA at 45 MINUTES! > >Here is a traceroute back to cdrom.com. > > 1 michnet2.msu.edu (35.8.2.3) 4.458 ms 2.598 ms 5.284 ms > 2 35.125.0.1 (35.125.0.1) 10.828 ms 4.955 ms 5.280 ms > 3 cpe2-fddi-1.Chicago.mci.net (192.203.195.5) 5.538 ms 5.952 ms 8.002 ms > 4 border1-hssi1-0.Chicago.mci.net (204.70.24.5) 19.377 ms 145.608 ms 75.289 > 5 core-fddi-0.Chicago.mci.net (204.70.2.81) 15.953 ms 11.500 ms 12.831 ms > 6 core1-hssi-2.NorthRoyalton.mci.net (204.70.1.94) 19.188 ms 21.190 ms > 7 core1-hssi-3.WestOrange.mci.net (204.70.1.102) 30.435 ms 35.809 ms > 8 core2-hssi-2.Washington.mci.net (204.70.1.106) 40.038 ms 46.752 ms > 9 borderx1-fddi-1.Washington.mci.net (204.70.74.52) 37.878 ms 40.587 ms * >10 mae-east-plusplus.Washington.mci.net (204.70.74.102) 236.610 ms 134.524 >11 mae-e-1.e0.crl.com (192.41.177.104) 50.809 ms 64.060 ms 77.006 ms >12 T1E-CRL-SFO-09-S0/0.US.CRL.NET (165.113.30.101) 172.132 ms 185.432 ms >13 F0-CRL-SFO-01-F0/0.US.CRL.NET (165.113.55.1) 144.838 ms 111.282 ms >14 wcarchive.cdrom.com (165.113.58.253) 97.066 ms * * > >Waitaminute, is that an ethernet I see between MCI and CRL? This is the second report we've gotten about this, and the last one had a nearly identical traceroute - MCI through mae-east. The last one I analyzed showed that the problem was the MCI -> MAE-east hop, not the MAE-east -> CRL hop. I don't know what to say about this other than it appears to me to be an MCI problem. CRL has future plans to upgrade their connection to MAE-east to DS3 (it is currently 10Mbits), but in this case it appears that it's not going to matter. :-( CRL peers with MCI at the Pacbell NAP (DS3), too, and in fact that is how the traffic gets back to you: traceroute to michnet2.msu.edu (35.8.2.3), 30 hops max, 40 byte packets 1 F0-CRL-SFO-01-F0/0.US.CRL.NET (165.113.58.1) 0.971 ms 0.727 ms 1.063 ms 2 pb.mci.net (198.32.128.12) 37.565 ms 95.140 ms 8.486 ms 3 borderx1-hssi3-0.SanFrancisco.mci.net (204.70.158.105) 4.853 ms 3.142 ms 4.223 ms 4 core2-fddi-0.SanFrancisco.mci.net (204.70.158.49) 5.521 ms 6.114 ms 8.334 ms 5 core1-hssi-2.Sacramento.mci.net (204.70.1.146) 12.456 ms 10.910 ms 9.344 ms 6 core1-hssi-2.WillowSprings.mci.net (204.70.1.174) 262.131 ms 223.868 ms * 7 core-hssi-3.Chicago.mci.net (204.70.1.85) 112.552 ms 172.846 ms * 8 border1-fddi-0.Chicago.mci.net (204.70.2.82) 143.960 ms 106.760 ms * 9 * * * 10 * merit-michnet-ds3.Chicago.mci.net (204.70.24.6) 180.087 ms 147.491 ms 11 fdd0.michnet1.mich.net (192.203.195.4) 135.588 ms 147.975 ms 145.068 ms 12 hssi0.msu2.mich.net (35.125.0.2) 145.792 ms * * It seems pretty clear to me by the above that the problem is with MCI, starting with the Sacremento->WillowSprings hop. -DG David Greenman Core-team/Principal Architect, The FreeBSD Project