6 messages in net.nether.puck.cisco-nsp[c-nsp] BGP hardware
FromSent OnAttachments
Ronan DalyJan 23, 2005 6:33 pm 
Elmar K. BinsJan 23, 2005 6:49 pm 
Simon Hamilton-WilkesJan 23, 2005 9:31 pm 
Brian FeenyJan 23, 2005 11:38 pm 
Gert DoeringJan 24, 2005 2:45 am 
Kristofer SigurdssonJan 24, 2005 2:55 am 
Actions with this message:
Paste this link in email or IM:
Paste this link in email or IM:
Atom feed for this thread
Paste this URL into your reader:
Subject:[c-nsp] BGP hardwareActions...
From:Elmar K. Bins (el@4ever.de)
Date:Jan 23, 2005 6:49:30 pm
List:net.nether.puck.cisco-nsp

Hi Ronan,

I'm looking into running eBGP with three providers and taking full tables from all three. I've currently got a 3725 with 256MB of RAM, a NM-2FE2W and a NM-1FE2W. Would that be enough to handle the three tables if I'm pushing under 20Mb/s?

When I started my current job 2 1/2 years ago I've been handed over a 3640/128 Meg, crappy configuration, with three full feeds (each about 110K prefixes then), pushing about 8 Mb/s. After cleaning the configuration, the box went from slow to ok, had a couple of CPU cycles to spare and never went down - it just became slow when we were hit by traffic storms.

I believe your 37 will hold it for the time being, but I wouldn't count on it keeping up in say two or three years' time (more prefixes, v6 tables, growth of your company, changes in traffic patterns...). Oh, and I'm not sure what the box will do with a couple thousand routes flapping; if you trust your upstreams/peers... ;-)

Yours, Elmi.

--

"Begehe nur nicht den Fehler, Meinung durch Sachverstand zu substituieren." (PLemken, <bu6o7e$e6v0p$2 at
ID-31.news.uni-berlin.de>)

--------------------------------------------------------------[ ELMI-RIPE ]---