[nsp] Preventing service degradation - automatic techniques

[Sam Stickland]

Hi,

I'm wondering what standby and routing strategies (if any) that people are
employing to stop the service degradation that occurs when a router doesn't
fail, but starts experiencing the sort of excessive CPU load that causes
packet loss (often caused by misconfiguration or unexpected IOS effects.

In these sort of circumstances HRSP doesn't change the active router because
it's still receiving hello's from the first router. Likewise the first
router would still be communicating with all or most of it's BGP peers, it's
just not doing a very good job with the traffic it gets given by them.

It's seems to me that it's pretty much only manual intervention that can
help in these circumstances, but I'd be intrigued to see what others might
have to say.

Sam