[c-nsp] strange RTT increase in ASR1006

Tassos Chatzithomaoglou achatz at forthnetgroup.gr
Wed Feb 20 14:00:30 EST 2013

Nick Hilliard wrote on 21/2/2013 01:30:
> On 20/02/2013 22:28, Tassos Chatzithomaoglou wrote:
>> If i knew that a single command can decrease throughput from 40G to 8G in a supposedly top
>> class cpu with everything hw-accelerated (aka qfp), i would have posted the nothing
>> special interface config... and would have looked at other platforms/solutions...from the
>> beginning.
> to be fair, nbar is kinda different.  40 gbit/sec of nbar classification
> with no performance hit would be a pretty impressive feat on any platform.
Miercom test says so for the ESP20...regardless of many others opinion (inc mine) about
their c-tailored tests.
Internal cisco tests talk about nbar performance of 10 x our current pps.
I guess we'll know for sure as we get an official answer from tac.
>> Anyway, previous releases didn't have this issue (at least with this amount of traffic),
>> so let's hope it doesn't become a new feature. I don't want to even think about future
>> nbar in ASR9k.:-\
> does the asr9k even do nbar?  I don't think it does even with the ISM module.
> Nick
Not that i know of. That's why i'm talking about "future" nbar

Many times ASR9k is being pushed as ASR1k's big brother, so i wouldn't be surprised if it
was to appear sometime in the future.


More information about the cisco-nsp mailing list