On 16/04/2010, at 5:12 PM, Dobbins, Roland wrote: > On Apr 16, 2010, at 1:35 PM, <[email protected]> wrote: > >> Now is there anything else that we need to consider while considering the >> support for netflow and the additional CPU load that will be introduced. > > NDE CPU utilization should be single-digit - while any feature can put a > router over the edge in terms of aggregate CPU, lab testing prior to > deployment with your anticipated feature-set and configuration, using > generated test traffic, should provide assurance and validation.
Obviously doing this in a lab as Roland suggests will provide better figures, but can anyone provide experience on throughput to expect using a 38xx or 39xx for basic BGP to 2-3 full table peers with the basic edge services (NetFlow, ACLs, uRPF) enabled, before its CPU hits the roof? A 3845 lists at 256Mbps and a 3945 at 502Mbps in the routing performance throughput tables, but I am interested in how much of that people are actually seeing... Regards, Tom _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
