[21680] in North American Network Operators' Group
Re: InternetWeek Article on UUnet: question
daemon@ATHENA.MIT.EDU (Sean Donelan)
Wed Nov 18 12:28:16 1998
Date: Wed, 18 Nov 1998 10:46:51 -0600
From: Sean Donelan <SEAN@SDG.DRA.COM>
To: gmatey@baynetworks.com, nanog@merit.edu
gmatey@BayNetworks.COM (George Matey) writes:
>Can someone from Uunet please comment on the information in InternetWeek
>and provide us (either on or off list) with more information so that we can
>fix any Bay issues encountered last Sunday?
http://www.internetwk.com/news1198/news111398-2.htm
I read the article and am puzzled by a number of quotes attributed
to the various individuals in the article. I don't know if Boyne,
Curran, Gaddis were just giving "the usual suspects" speeches, or
had concrete evidence to back their statements.
Was NETCOM identified as the source? Why was the NETCOM NOC difficult
for UUNET to reach on Sunday? Was it a 'customer configuration error'
or a router vendor(s) software bug? If UUNET had identified the correct
customer router to block, why did the route return? Why was SAVVIS's
AS number in some of the traces if Savvis had shutdown its connections
to prevent the bad route from propagating? Did Savvis customers' really
suffer no ill effects, and routing the traffic through C&W and Sprint
restored all service, or are 'benchmark guarantees' measuring something
else? Is Mr. Curran correct that there is really nothing a telephone
end-user can do to harm the telephone network when using out-of-spec
equipment? Is Mr. Curran's solution of the IETF and IOPS to work on
filtering customer router information the correct solution, or should
router vendor's implementations of the protocols undergo more extensive
inter-operability testing before being put in service?
Don't you hate this Andy Rooney routine?
--
Sean Donelan, Data Research Associates, Inc, St. Louis, MO
Affiliation given for identification not representation
