[c-nsp] 6500 Netflow

[Tim Durack]

On Fri, Apr 18, 2008 at 8:55 AM, Gert Doering <gert at greenie.muc.de> wrote:
> Hi,
>
>
>  On Fri, Apr 18, 2008 at 08:39:53AM -0400, Tim Durack wrote:
>  > Sounds like an argument for doing something stateless like sFlow for
>  > statistics instead of flogging the NetFlow horse.
>
>  This very much depends on your needs.  If you need detailed records for
>  billing purposes, sFlow (or sampled NetFlow) is just not good enough.
>
>  For basic "what's going on?" detection, sflow/sampled netflow is fine.

Yup, interface counters for billing, sFlow for everything else (I
agree that probably doesn't work for everybody. We manage to collect
data on 12k switch ports using XRMON/sFlow and Inmon's Traffic
Sentinel without too much trouble.)

Cisco's argument against sFlow seems to be: it's not as accurate as Netflow.
Then they introduced sampled netflow because scaling netflow on router
hardware is challenging.

I find that reasoning hard to follow.

Tim:>