[108575] in North American Network Operators' Group
Re: Help needed - Cisco Netflow
daemon@ATHENA.MIT.EDU (Phil Bedard)
Fri Oct 10 14:17:43 2008
From: Phil Bedard <bedard.phil@gmail.com>
To: "Lee, Steven (NSG Malaysia)" <kin-wei.lee@hp.com>
In-Reply-To: <084962C061414240A0CDB4BE328A9B2D097FCBC294@GVW1100EXC.americas.hpqcorp.net>
Date: Fri, 10 Oct 2008 14:17:18 -0400
Cc: Julio Arruda <jarruda-cnsp@jarruda.com>,
"nanog@nanog.org" <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
It depends on how many active flows you have at any one time. Also, I
don't think the SIP-601 supports full netflow V5 in hardware, only V8,
which is aggregated netflow, which may not be what you want. It does
do V5/V9 sampled netflow in hardware. The sampled netflow on that
platform is better than on say the 6500/7600, so I would just go with
that.
Phil
On Oct 10, 2008, at 1:20 PM, Lee, Steven (NSG Malaysia) wrote:
> Hi all, the LC on 12000 series is 12000-SIP-601 (engine 5) with
> multiple STM-1 SPA card. As far as I know the engine 5 card cannot
> do 450Kpps on netflow. Since the STM-1 is mainly for NB-AMR traffic
> (95%), do you recommend the sampled netflow instead? Does anyone
> aware of the sampled netflow accuracy?
>
> -----Original Message-----
> From: Julio Arruda [mailto:jarruda-cnsp@jarruda.com]
> Sent: Saturday, October 11, 2008 12:43 AM
> To: Lee, Steven (NSG Malaysia)
> Subject: Re: Help needed - Cisco Netflow
>
> Lee, Steven (NSG Malaysia) wrote:
>> Hi all, I have a customer who has a MPLS network for E/// Media
>> Gateway (MGW) NB-AMR VoIP traffic. The packet size for the NB-AMR
>> traffic is fixed size 110 bytes. During the high load period, it
>> can reaches 450Kpps on a STM-1 link. The router platform is Cisco
>> 12000 series, can the router generate 100% netflow data with such
>> traffic load? Also can someone suggest a carrier class netflow
>> collector for mobile service provider environment.
> Just a quick note..
> You may want to post the Line Cards they have...Distinct LC have
> distinct Netflow capabilities (and impact).
>
