|
I concur with ChuckMountatin - if you want to test the performance of a networked device, don't use Internet based speedtest site - use local resources such as NetIO or iPerf - both free and cross platform. With these tools you essentially create your own "speedtest" server within your local environment and thusly don't have anything out on the Internet muddying the water.
As I am fond of saying, these speedtest tools don't actually test the "speed" of anything: They send a measured amount of data, time how long it takes and compute an average. It's kind of like the speedometer in a car versus trip computer. It's a useful enough indicator, but pros' wouldn't be getting in a twist because the speedtest said one number and the "link rate" is another (the two are not at all the same thing in any case.) So, if one were testing a Gig ethernet NIC and were getting back iperf/speedtest numbers in the 800-900 range, we'd be happy - we wouldn't be worrying that is wasn't 1000. But if it was 80-90 (ie an order of magnitude lower) we know something was "up" - like there's a 100mbps link in the pathway somewhere.
I have seen instance where rather than group policy settings nobbling the rates, some on-machine firewall and/or proxy can clobber the rates also. Again, something else to check out with your company's ITD. |
|