Determining which multicast client is the bottleneck?

  • We have 1 Gbps connections to most of our computers, and we often image labs of 25 to 45 computers. Very often, one or two computers will hold up the rest, either because of a slow hard drive or (more likely) a slow network connection. Even though a connection might technically negotiate at 1 Gbps, the actual throughput is often much lower.

    While multicasting, is there anything in FOG that will give tell you which computer is causing the issue? We’re using 1.2.0.

  • I get what you’re saying. Nonetheless, I would hate to return to unicast because in the rooms where everything works correctly, I can image 35 computers with a 42 GB image in less than 8 minutes. That’s really hard to beat. It’s getting so fast that we can just about use FOG to push out an image whenever we need even small changes made to the room and just forget about using VDI, as that is VDI’s big advantage (i.e., being able to quickly push out changes to a whole room of computers without much hassle).

    We could go computer-by-computer to find the culprit, but that’s such a pain. That would be the perfect feature for FOG, as it would know better than any other tool which computer is being the hold-up.

  • Moderator

    no, I believe not.

    I personally prefer and stick to unicasting as I have experienced as you have that a dodgy network connection (cable/nic) or a dying HD can bring the process to a halt.

    With the ability of FOG to que the rest of a group and the speed when unicasting a lab I can set a unicast job per room and not have to keep an eye on the process like I did when trialing multicasting a few years back.