We used PerformanceTest 6.1 by PassMark which you can find out more information about here. It has a handy Advanced Networking section which is perfect for our testing.
For our testing of the NETGEAR GS605AV we did things a little different. I set up a NAS appliance with a few HD encoded movies (MKV files at 1080p) and then connected my media playback system to different ports on the GS605AV. At the same time I also threw in a large (6.5GB) file transfer to see what the results would be.
In the graph below you see best possible speeds recorded using PassMark. However, we also wanted to gauge real world use to see how the different pre-set levels of QoS would work.
- Raw Bandwidth
As everything was hardwired, you can see the speeds in our best possible setup testing are very good. They are what you would expect a gigabit switch to be able to give you.
However, we found some interesting items once we started our real-world test. The GS605AV worked flawlessly with the Seagate BlackArmor NAS-220 as the media storage device (plugged into the dark green port) and our two "work systems" plugged into the two light green ports. I was very impressed with the way it handled traffic. As soon as I started up a movie the file transfer slowed down to allow smooth playback of the video I had selected. This is exactly what you would expect from a switch with a QoS function and shows that the QoS levels are setup pretty well.
When I began to move things around I noted different results (of course) and encountered an issue. With the Download system in the dark green port and the NAS and Media Playback system in the light green I saw both media and file transfer impacted. I started to note stuttering and the occasional jump in the MKV file while the file transfer also slowed down. The reduction in speed was not as large as when the download system was in the light green port, but it was still noticeable.
As you would expect, when I moved things down to the yellow ports I began to have more issues when I loaded up the switch with traffic. If I was doing one or the other, everything was fine no matter where the items were plugged in. It was only when I started up other items that we saw the QoS kick in, which is as it should be.