[c-nsp] How much will a service policy kill a router?

[Rodney Dunn]

Yes. They did some improvements to try and bring that performance
back inline to what it was before they added the new classification
infrastructure in 12.4(11)T.

I've not seen or done comparable testing to know how much difference
there is.

Rodney


On Thu, May 31, 2007 at 09:21:05AM +0300, Hank Nussbacher wrote:
> At 03:35 PM 30-05-07 -0400, Rodney Dunn wrote:
> >Impossible to answer without knowing the full classification model
> >you use.
> >
> >thousdand line acl's...hit is a lot.
> >
> >short acl's .... smaller.
> >
> >
> >12.4(11)T ...large hit until we fix some of the problems we caused
> >with a new classification infrastructure.
> 
> We opened a TAC case on this a few months back and were recently told that 
> 12.4(15)T that is due out end of June will be better than 12.4(11) in 
> regards to QoS but not as good as 12.4(4)T was.
> 
> -Hank Nussbacher
> http://www.interall.co.il