Huge WAN spike when VEP backing up to remote repository

[mbrownwfp]

We are transitioning to VEP backups for our remote office locations. These sites have a single file/print server so the basic VEP backup model is sufficient for our needs. We are backing up to a central Veeam repository located in our primary datacenter location.

The remote sites have various speeds of WAN links, most are on symmetrical 10Mb or 100Mb fiber, but some of the more remote sites are still on T1 or even asymmetrical ADSL.

We have seeded Full backups for the slow sites into the central repository using portable storage but are now running into issues with massive latency on the slow WAN links while incremental jobs are running.

I've enabled the Throttling option for the remote subnet on our VBR server that manages the central repository. As the VEP job is running it does indicate that throttling is enabled however latency on the site's WAN link jumps to 2000ms and beyond as soon as the job starts processing data.

Anyone else run into this before? Any workarounds available?

This was not an issue with our previous backup software (Symantec NetBackup with client-side dedupe enabled).

[jmmarton]

What I've recommended to customers needing to backup physical Windows servers at remote sites is to add storage to each server if possible (even a USB drive) and configure each server as a repository withing Veeam Backup & Replication. That way VEB/VAW backups can be performed to a local repository, then VBR backup copy jobs can be used to get the backups offsite. For sites with slow WAN links you can then leverage WAN acceleration with those remote sites.

Joe