Ghost Solution Suite

Hey all....I have to image classrooms on a regular basis, and until recently I have been pleased with the transfer rate of the images. The network is a 1 GB network (cabling,  unmanaged switches, and NICS). I would constantly get +1500 MB transfer rate to 20 machines at a time. However, over the past few days, my transfer rate has dropped to barly holding at 200 MBs...and sometimes much lower than that (around 60-70 MBs). Machines vary, but are all Dell 745s, 755s, 780's, and 790s. Drivers are either Broadcom 5XX for the 745s or the Intel 1000 for the others. Boot straight into Ghost from the console using WinPE.

Nothing has changed. Im green across the board for the 1 GB network, and I cant see any broadcast storm happening unless im multicasting. Im at a loss as to why the sudden slowdown, and have no idea where to check first. I've tried multiple images, so its not a problem with a specific image or room. I did notice however that since they are unmanaged switches, the image im sending is going EVERYWHERE. If I remove the other rooms from the network, the transfer rate does increase slightly, but not by a large amount.

Any ideas on where to check? I see alot of people posting they get 200-300 MBs, so I guess I got spoiled seeing +1500 MB. I want that back as alot of these images are 70-80 gigs each and I have to image them on a weekly basis!

I know, I know..I could solve it all with a nice VLAN switch, but tell that to my boss :)  Hope someone has some ideas as to why the sudden slowdown...Thanks for all your help!

The first and obvious question to ask is: "What has changed in the last few days?"

We are a week past "patch tuesday" so has your 'server' machine received any updates in that time which may have affected the network?

Also, have you tried systematically removing machines from the group of 20 to see if there is an issue caused by a single NIC that is affected the rest of the network.  Isolate your 20 machines, then unplug the NICs on half of them and test. Then try the other half and retest.  This will quickly reveal whether any one machine is dragging the rest down.