We are considering backing up from one site to another site directly, over a 10Mbps WAN connection.
So we are not going to use replication, just a backup job. So VM's from site A will be stored at storage on site B.
We would like to use the forever incremental option for this so we don't have to push a full backup every week.
As I understand, the forever incremental process has to pull the changed blocks from the full backup and store it in VBR files, than it pushes the newest changes into the full backup file. These would be a lot of reads and writes over a 10Mbps WAN connection.
So my question is, where does this processing take place? Am I correct in thinking that the Backup Proxy actually does ALL the work? So it will read the data and also do the processing for the forever incremental?
In which case my concept won't work, because
A) I either have the proxy at site A where it can read the VM's, compress them and send to the repository, but also has to read - OVER THE WAN - from the repository to make the VBR files.
OR
B ) I place the proxy at site B in which case it has to read the VM's over the WAN in uncompressed format..
Am I correct, or not?, or am I missing an alternative
Thanks in advance for any thoughts on this