[nsp] 7206vxr pps & mbit/s limits

jlewis at lewis.org jlewis at lewis.org
Mon Oct 27 19:33:20 EST 2003


I see there was a similar thread about a year ago...but now I'm wondering 
what I should reasonably expect from a 7206vxr NPE300.  I've got one 
routing betweeen several FE's, doing OSPF, BGP, ISL, some ACL's, policy 
routing to block nachi pings.  CPU load is pretty constant around 50% 
except for the bgp scanner spikes.  

We recently moved a FE transit connection to it, and that seems to have 
pushed it over (or at least a lot closer to) the edge.  If I add up all 
the 30s avg bps/pps, I get (at this moment) 125mbit/s and 37kpps.  When we 
moved that busy transit FE onto this router, we immediately started 
getting fairly constant underrun errors on the interface and output drops.  
I was able to considerably reduce the output drops by increasing the 
hold-queue, and this may have even reduced the frequency of underruns, but 
we still get them regularly.

Previously, this transit FE was on a 7500 (RSP4, VIP2-50s) that seemed 
perfectly happy with the traffic.

Is upgrading to a faster NPE (NPE400?) going to make much difference, or 
would we have to go to NSE1 or NPEG1 to see much improvement?  I'm 
wondering if we ought to just go with another 7500...perhaps the 
distributed architecture really does beat the 7206 and its single 
processor.

----------------------------------------------------------------------
 Jon Lewis *jlewis at lewis.org*|  I route
 Senior Network Engineer     |  therefore you are
 Atlantic Net                |  
_________ http://www.lewis.org/~jlewis/pgp for PGP public key_________



More information about the cisco-nsp mailing list