[net-perf] Getting Started?

[Michael Sinatra]

>>1) What performance/connectivity tools have you deployed?

iperf (see http://net.berkeley.edu/perf/)
smokeping (currently in "proof-of-concept" stage)
NLANR AMP
CSG/Surveyor (broken and possibly defunct)

>>2) Are they publicly accessible?

iperf - yes
smokeping - no (need to move it to a production server)
NLANR AMP - yes
CSG/Surveyor - theoretically yes, but their website is broken

>>3) What tools are you planning on deploying?

ORNL/ANL java b/w tester.  Had done a POC, but took the old machine 
offline and hadn't gotten around to re-doing it for production.

Also, I think we have funding for a cluster of test boxes to deploy on 
campus.  I am interested in ideas (AMPlets, etc.) that this group has 
for using them.

>>4) What should be the minimal set of tools each campus should have deployed?

I'd like to see bwctl, a traceroute server (web based, as in 
<http://net.berkeley.edu/cgi-bin/traceroute>), possibly a multicast 
beacon, maybe an AMPlet?  But iperf at a minimum.

>>5) What is the minimal set that should be publicly accessible?

Iperf, AMP-type reports, maybe smokeping, traceroute.  I agree with 
MVN's comments re limiting access (which is why bwctl is appealing).

michael