spikespiegal, your personal example may serve you well. In small VMware clusters you can often do ok with gig and a couple of hosts, 15-20 and vmotion on Gig. I have no problem doing that kind of stuff on my test environment. However when you have 1 host with 8 sockets and 6 cpu/s socket + few hundred gig of RAM you need 10GIG performance even in vmotion. If you need to move 512GB of RAM over to another machine (host failed or is going to maintenance mode) you need (in theory, no overhead included) 8.2 seconds per gig of RAM moved on a 1 gig connection. You are pushing nearly 70 minutes to vmotion that host. You may not have that time in an HA event. (vmotion is a good example lots of time because there is a lot of RAM to network to RAM movement so disk means nothing) You won't see this at the small and even medium level business. Data centers however do this regularly. 10gig would do the entire transfer in 7 minutes.
Other places I have seen it is in the financial markets moving stock symbols around.
Other places I have seen it is in the financial markets moving stock symbols around.
