Date: Tue, 12 Mar 1996 08:15:32 -0800 From: David Greenman <davidg@Root.COM> To: John Curran <jcurran@bbnplanet.com> Cc: mark@grondar.za, chat@freebsd.org Subject: Re: Problems in SF? Message-ID: <199603121615.IAA21042@Root.COM> In-Reply-To: Your message of "Tue, 12 Mar 1996 10:42:59 EST." <v02130500ad6b49d61969@[199.94.220.15]>
next in thread | previous in thread | raw e-mail | index | archive | help
>Mark, > > A traceroute would be most appreciated. Please email to myself > or "ops@bbnplanet.com" (as I can't be certain that everyone is > FreeBSD followers...) > >/John Hi John. I suppose since I'm one of the ones complaining the loudest and since I'm one of the people responsible for these problems in the FreeBSD Project and at WC CDROM, I feel obligated to respond. Remember a few weeks back when you pointed out to me in email that BBN Planet had T3's going in all directions from Stanford and I was saying that it didn't look that way to me? - that everything went through the microwave link to MCI? Well, part of the problem people are complaining about now is that the traffic is taking a much more indirect path (no longer through MCI), going across the US and back: [freefall:davidg] traceroute implode.root.com traceroute to implode.root.com (198.145.90.17), 30 hops max, 40 byte packets 1 gatekeeper.cdrom.com (192.216.222.1) 1.887 ms 1.908 ms 1.885 ms 2 ucb-o2.bbnplanet.net (131.119.70.133) 3.174 ms 3.389 ms 3.130 ms 3 ucb-2.bbnplanet.net (131.119.253.5) 4.077 ms 4.231 ms 6.733 ms 4 * su-tk.bbnplanet.net (131.119.2.1) 7.045 ms * 5 su-b.bbnplanet.net (192.31.48.66) 12.307 ms 11.847 ms 9.976 ms 6 4.0.1.2 (4.0.1.2) 58.903 ms 60.336 ms 62.462 ms 7 4.0.1.6 (4.0.1.6) 75.344 ms 74.071 ms 73.897 ms 8 icm-fix-e-F0.icp.net (192.203.229.245) 82.637 ms 78.055 ms 74.084 ms 9 icm-dc-2b-H3/0-T3.icp.net (192.157.65.121) 76.889 ms 76.954 ms 78.658 ms 10 sl-dc-6-F0/0.sprintlink.net (144.228.20.6) 74.342 ms 204.247 ms 163.218 ms 11 sl-stk-5-H1/0-T3.sprintlink.net (144.228.10.2) 165.794 ms 181.430 ms 175.849 ms 12 sl-stk-11-F0/0.sprintlink.net (144.228.40.11) 166.467 ms 175.594 ms * 13 sl-intworks-1-S1-T1.sprintlink.net (144.228.141.66) 171.641 ms 162.277 ms 168.508 ms 14 implode.Root.COM (198.145.90.17) 206.437 ms 199.045 ms 191.234 ms The round-about path to Sprint is a new problem that started last week. I'm assuming that this is caused by BBN Planet not peering with Sprint in the Bay Area (at the PacBell NAP or MAE-west) coupled with some changes in the routing tables to prefer BBN Planet's network. The above is compounded by the older problem that *appears* to be between ucb-o2 and ucb-2 (see above). The lossage in this particular traceback makes it look like the problem is Stanford<->Berkeley - and it might be - but I did a variety of 'ping' tests (yes, I know that pings are a poor way to look for packet lossage, but it's the only tool I have for this) about 5 weeks ago and it appeared that I could reliably ping ucb-o2, but get 10-15% packet loss when pinging ucb-2 (this is from WC CDROM). Gary Palmer of Walnut Creek CDROM filed a trouble ticket about this problem and except for the ticket-filing confirmation, we never heard anything more. When we arranged for CRL Network Services to provide the 100Mbps colocation of wcarchive.cdrom.com, we also arranged as part of a package deal for CRL to provide T1 services to WC CDROM in Concord. We've been waiting for MFS to finish the installation of the new T1 and will be moving over to that as soon as it is ready (starting tomorrow if everything goes right). -DG David Greenman Core-team/Principal Architect, The FreeBSD Project
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603121615.IAA21042>