Hello, I can't resist jumping into this disscussion. I think its clear that there is much work to be done on router/IP performance. The question then is what can be done in the meantime, until said work is completed? Here is a script I wrote that will work on cisco routers. The script takes advantage of cisco's enterprise specific MIB. There are 3 SNMP variables yielding CPU busy percentage. The first is CPU busy percentage in the last 5 seconds of scheduler time. The other two are 1 and 5 minute exponentially-decayed moving averages. The CPU busy metrics can yield interesting results, for example %'s above 40% indicate a busy router. I was able to generate some interesting graphs awhile back for our cisco routers. For more information, use the oid's in the script as pointers into the cisco mib. And here is the script for anyone who is interested. --Gerald Winters <gerald@merit.edu> #!/usr/local/bin/perl @mib_objs = ("1.3.6.1.4.1.9.2.1.56.0", "1.3.6.1.4.1.9.2.1.57.0", "1.3.6.1.4.1.9.2.1.58.0"); for ($i = $num_values = 0; $i < 3; $i++) { foreach (`/afs/umich.edu/user/g/e/gerald/CISCO/snmpget -h 141.211.128.1 @mib_objs[$i]`) { if (/value:/) { $num_values++; @values[$i] = /\w+\W+\w+\W+(\w+)/; } } } if ($num_values == 3) { print "@values[0] @values[1] @values[2] ".`date`; } else { print "0 0 0 ".`date`; }