[c-nsp] CPU utilization - "media converter" vs "bump in the cable"

[Rick Ernst]

I was a bit surprised to see that a 7206VXR/NPE-G1 running at the same CPU
utilization on both an ethernet upstream with ~300mbs (in+out) running
through it and an OC-3 upstream with about 100mbs through it.

Multiple upstreams for the same ASN, essentially the same configuration
(other than IP addresses).  All running BGP with full tables to different
upstreams.

Is the NPE-G1 very non-linear in CPU load or (my guess) that the OC-3
upstreams are spending more time rewriting the packet headers from OC-3 to
ethernet.

I'd like to better understand what is going on for capacity/scalability
planning.

Thanks,
Rick