Date: Wed, 24 Jun 1998 20:48:26 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: wpaul@skynet.ctr.columbia.edu (Bill Paul) Cc: tlambert@primenet.com, opsys@mail.webspan.net, jkh@time.cdrom.com, hasty@rah.star-gate.com, hackers@FreeBSD.ORG Subject: Re: option MROUTING Message-ID: <199806242048.NAA14440@usr04.primenet.com> In-Reply-To: <199806241249.IAA07653@skynet.ctr.columbia.edu> from "Bill Paul" at Jun 24, 98 08:49:55 am
next in thread | previous in thread | raw e-mail | index | archive | help
> Then maybe he can explain why Columbia's MBONE service from Sprint > sucks so much. Traffic from Columbia goes from our primary gateway to > mbone.appliedtheory.com, and then to pen-mbone-1.sprintlink.net, > which, I've been told, is a Sun SPARC machine of some kind running > Solaris. It has a large bunch of tunnels hung off it, and it leaks > packets like a sieve. When connectivity isn't screwed up like it is > now (more on this in a second), it loses anywhere from 40% to 60% of > all the traffic passing through it. The shuttle mission astronauts > can hear NASA better than we can when a mission is being multicast. Have you made the appropriate QOS requests at the link layer for what is, in fact an ATM transport which is intentionally using the standard ATM "leaky bucket" packet forwarding algorithms? If not, it's now wonder that your bucket is leaking. However, I would be much more likely to blame the fact that the Gigaswitches at NASA Ames are insufficiently interconnected because some manager decided that there was no need to buy two more instead of one more when they needed more capacity than one could handle (can you say "flunked college algebra because of a failure to understand combinatorics? I knew you could..."). Try going someplace that doesn't depend on NASA, and posting a comparative analysis for us. PS: Why don't you contact Vadim and ask him; you are probably right that he can explain your problem to you... PPS: I don't like Sprint. They are cheap (not inexpensive) and they failed to upgrade their routers to get around the route drop problem when the rest of the world did; it took them a long time to come up to par. That said, I have great respect for Vadim's work, expescially that in BSDi and teh stuff that made it into FreeBSD through CSRG. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806242048.NAA14440>