>From: Avi Freedman <avi@freedman.net>
>To: rbuchals@hotmail.com (Ron Buchalski)
>CC: avi@freedman.net, simonl@rd.bbc.co.uk, cisco-nsp@puck.nether.net
>Subject: Re: [nsp] RSP/VIP performance question
>Date: Fri, 10 Sep 1999 15:10:15 -0400 (EDT)
>
> > >I'm sure you're not, but that's not what we've seen with a normal mix
> > >of production traffic, is all, across many different IOS versions, with
> > >cef, optimum, flow, and distributed or not.
> >
> > How do you know that the limits you were hitting were a result of packet
> > switching limits or bandwidth limits on the VIP2-40 versus something
>else,
> > either internal to the router (other processes eating up CPU) or
>something
> > external to the router (net congestion)?
>
>Nothing else was particularly loaded (cpu/bus/...), lots of configs were
>tried, lots of different hardware, and we've never been able to achieve
>anything like 400mb out of a VIP 2/40.
>
I didn't say that you could get 400mb/s out a a VIP2-40.
> > >Are those reality-numbers or theory-numbers?
> >
> > As I said, they are raw performance numbers. Depending on your
> > configuration, topology, network conditions, etc., YMMV.
> >
> > Better stated: You may see worse performance than this, but you won't
>see
> > better performance than this.
>
>Well, that's dancing around my question a bit.
I wasn't intending to dance around the issue.
>
>The question was: are those raw-performance numbers based on observation
>on a real router, or just theoretically computed?
No, I haven't observed the numbers. I'm basing it on the numbers used in
the configuration guidelines, which are based on design goals which the
product can achieve.
Also, keep in mind that the number (300mb/s, 360 mb/s, whatever) is _total_
throughput (input plus output). So, line rate can be achieved with a POS PA
(155mb/s * 2 or 310mb/s). Of course, Layer 1 and Layer 2 overhead make the
actual Layer 3 throughput less than 310mb/s.
-rb
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:05 EDT