[c-nsp] 3640 CPU load question

[Burton Windle]

I have a 3640 with 128mb memory running 12.2(24). It has one FastEth
interface, and 8 serial interfaces (each with an Adtran T1 CSU hanging off
of them).  The network setup is hub-n-spoke, with the 3640 being the
center of the WAN hub.

I would think the router should be snoozing; the FastE int pushes about
1.5 megabits/sec (700 packets/second), and the busiest T1 averages much
less than a full T1 (200pps, 400kilobit/sec). I have no real ACLs to speak
of, the router doesn't do any NAT, and runs EIGRP. I do MRTG polling for
the interface bandwith stats every 5 minutes, and the router exports
NetFlow data. I also have 'ip tcp header-compression' & 'ip rtp
header-compression' turned on for each serial interface.

However, the router sits at 25% CPU usage (CPU utilization for five
seconds: 30%/30%; one minute:  25%; five minutes: 25%) all the time. Does
this CPU usage seem high?

A 'show proc cpu' doesn't show anything interesting:

5-3640-WAN#show proc cpu | exclude 0.00
CPU utilization for five seconds: 30%/27%; one minute: 28%; five minutes: 27%
 PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
   3        7768      1233       6300  2.03%  0.61%  0.44% 130 Virtual Exec
  21     1735868   4072677        426  0.08%  0.06%  0.03%   0 Per-Second Jobs
  31     5343616  13613338        392  0.08%  0.13%  0.14%   0 IP Input
  70     2664044  19148167        139  0.08%  0.09%  0.06%   0 IP-EIGRP Hello

Here is a snippet of the config for the interfaces:

interface FastEthernet0/0
 ip address 192.168.1.4 255.255.255.0
 ip route-cache flow
 speed 100
 full-duplex
!
interface Serial1/0
 bandwidth 1536
 ip address 10.255.110.1 255.255.255.252
 service-policy output WAN
 ip tcp header-compression iphc-format
 serial restart-delay 0
 ip rtp header-compression iphc-format
!



A gentle proding with a Clue-By-Four would be appreciated.

-- 
Burton Windle                           bwindle at fint.org