3
I was wondering if anyone could shed some light on implementing round robin NIC teaming, considerations, and best practices. Looking for some concise and clear information regarding it.
I do a lot of drive imaging over the network and saving time during drive imaging would save us a lot of money. Our biggest throttle so far has been our network. We are currently limited by a gigabit infrastructure.
I would like to implement an imaging station running some flavor of *nix. This imaging station would ultimately connect to a SAN volume over NFS. My thought was I could pick up some NICs and throw them into the imaging station. Then get another computer with the same amount of NICs and connect them directly over a bonded NIC team utilizing round robin scheduling. I believe between the imaging machine and the other machine we would in theory have connection able to achieve through-put considerably higher than that of a single Gbit connection. Yes?
In summation, would attempting to team 4 NICs and using round robin scheduling in a point-to-point connection between two computers achieve throughput in a single data transfer at or over 3 Gbit/s?
are you transferring the same image? Are you multicasting? – Keltari – 2013-05-07T18:15:29.327
@Keltari Single transfers of unique disks. This is not an imaging environment for client machines. One disk at a time, a different disk, every time. – 0xhughes – 2013-05-07T18:34:41.800