Switch/Router testing

You can use SiSoft's Sandra to test network throughput. For the hubs (and the switches, for that matter), I'd suggest running two tests on each, one with only two computers hooked up, and one with all the ports full (if you have enough computers to pull it off.
It'll give you more accurate tests. You can run tests with varying amounts of computers in between, but that's probably just overkill.

-----------------
Whoever thinks up a good sig for me gets a prize

what capacity are your trying to test and how fast are they supposed to be?

You'll need at least 4 pcs/servers with NICS rated to at least the speed of the interfaces on the switch and router.

Test for router -

setup a very large file transfer over it between tow hosts, note speed transfer runs. DOS FTP client will give you tatal bytes transferred and time taken for example.

then do it again, then setup a second identical transfer from the other two machines over the router. Check that the test only takes twice the time of the first, not 4 or 6 times, in which case you have buffer/memory/packet drop problems and have either setup the router incorrectly or it is cheap junk. The type of router will greatly effect its performance so I cannot really give you an estimate on that but a 10Mbs router with no IP filtering should sustain a healthy 3-4Mbs load.

For the switch do much the same, but in that case there should be no difference in running 1 transfer or 2 between different pairs. If there is, then the switching fabric is junk, or the switch is switching in software and therefore only useful in today's world as a doorstop. You should see little difference between a switched transfer time and the transfer time of the two pcs connected by a crossover ethernet cable.

good luck.

-* This Space For Rent *-
email for application details