Enron Mail

From:philippe.bibi@enron.com
To:mark.frevert@enron.com, greg.whalley@enron.com
Subject:Network
Cc:
Bcc:
Date:Tue, 16 Jan 2001 00:28:00 -0800 (PST)

FYI

I will keep you posted on the progress....



---------------------- Forwarded by Philippe A Bibi/HOU/ECT on 01/16/2001
08:26 AM ---------------------------


Mark Pickering
01/15/2001 03:26 AM
To: John Sherriff/LON/ECT@ECT
cc: Philippe A Bibi/HOU/ECT@ECT, Michael R Brown/LON/ECT@ECT
Subject: Network

Hi John,

I assume you spoke with Nigel following your voicemail on Friday and he has
brought you up to date.

We are doing everything we can to make fix this problem, and we are trying to
ensure everyone is in the loop.

The problem looks to be a software bug somewhere in the Cisco operating
environment. This has been escalated to Cicso's highest priority and we have
had up to 6 Cisco engineers working on the problem. We have also flown in one
on the Houston Netwrok team to assist. The problem we have is that the two
core 6500 routers, which are at the hub of the network will run along
perfectly at about 7% utilization, then, for an as yet undetermined reason
they suddenly jump to 99% and start thrashing, dropping network connectivity.

The catch 22 we have is that when the system goes down, the Cicso guys need
time to capture diagnostic information, but of course we just need to get it
up again straight away.

The team have systematically gone though each hardware component and each
software component to ensure all are working, all are at the latest cisco
recomended release level and all are required. Those not needed are removed
to ensure the 'simplest' environment possible.

The bottom line is that we still haven't solved the problem, but are doing
everything in our power to resolve it ASAP.

Regards
MRP

fyi, we have been working on a network redesign since before these problems
start which will distribute the network across more Core routers, (in effect
seperate networks). This design is the same as that proposed for the new
Houston building. Whatever design is in place, however, we can never rule out
a problem due to faulty software such as this.