Feature request: WAN acceleration bypass

[mikeschultzca]

Lets face it WAN acceleration is slow when you have a lot of VMs and not always necessary for all of them. It would be great if there was a way to bypass the WAN accelerators for select VMs and put them into direct mode. Bonus points if Veeam could do this on its own with size thresholds.

Currently each VM in my environments takes 10 minutes to process regardless of size, and because WAN Accelerators only process one VM at a time my only recourse is to build more jobs to speed things up and make use of my hardware.

[Vitaliy S.]

Thank you for the FR, however I guess you can configure a separate job for VMs that you don't want to run through the WAN accelerator? Would that be a good workaround?

[mikeschultzca]

Currently that is the only work around, but it would be nice to be able to able to keep it in a single job if possible. Also it can be hard to tell which VMs to move out of one job and into an other job without running it the job to find out how much change data each VM has.

[Vitaliy S.]

That's where I'm pointing to - if it is hard for a human to understand which VMs should be skipped, how WAN accelerator should do it? Do you have any suggestions for this?

[dellock6]

I believe the only parameter should be the processing time, and to figure out this value you have to run an A/B test against the same VM, with and without wan acceleration. The faster of the two tells you if wan is going to improve performance.
And to the original point Michael, wan-x is designed to trade disk for bandwidth, so it's not about being slow, but more about saving bandwidth at the expense of some other resource. The final result it's not changing, but wanted to point this out...

[mikeschultzca]

If it was possible for Veeam to scan the CBT database and take a guess at the amount of change data for a vm that would allow veeam to make the determination on its own. But a simple exclude button in the data transfer window would be all am I looking for right now.