[c-nsp] High CPU due to BGP Router process after changing BGP neighbor descriptions

[Alex Rubenstein]

On Mon, 25 Jul 2005, Clinton Work wrote:

> After making the description changes the CPU on the GRP was pegged at 99%
> for over 15 minutes. The GSR is running 12.0(25)S3.


<rant>

Perhaps not the best forum for this, but..

Why is it the cisco has code for BGP what is horrendously inefficient, in 
terms of CPU utilization? 7200's and 7500's and 6500's regularly sit at 7 
to 15% CPU simply handing BGP Scanner with only one full peer?

How is it that other vendors don't have this problem? What is so 
fundamentally different in IOS than that of JunOS or whatever? A Juniper 
RE-333 can handle routing updates with ease, yet the RISC's in the cisco's 
regularly keel over. I've seen 6500's with two full views take upwards of 
15 minutes to fully converge.

I am not cisco bashing -- I am just looking for an answer.

</rant>


-- 
Alex Rubenstein, AR97, K2AHR, alex at nac.net, latency, Al Reuben
Net Access Corporation, 800-NET-ME-36, http://www.nac.net