Questions on Off-site backups

[hollow]

I am just left with configuring off site backup to close my implementation of Veeam 7 project!

Scenario:
Site A Source
1 x VBR Master (For backup jobs only)
1 x physical server proxy and repository

Site B Destination
1 x VBR Master, also the proxy as well as the repository (For Replication jobs only)
1 x VM for offsite backup.

I understand that there is a copy backup job function with WAN acceleration enabled on both source and target.

Question:
1. When the copy backup job runs, which resources are the copy backup jobs using?
a. Source Master or b. Source proxy or c. Destination VBR master (also the proxy) or d. Destination VM for offsite backup?

I am guessing it is b. Source proxy and c. Destination VBR master (also the proxy)

2. If I am correct, this leads to my next question:
If the VM for off site backup does not require any resources (CPU/RAM) for the copy job, is 2 vCPU and 4GB RAM suffice?

3. Last but not least, is there any rule of thumb on the resources required vs the amount of copy jobs performed?
Eg: 1vCPU 1GB RAM per VM on both source and target?

Thanks in advance!



Please help me out. Thank you

[foggy]

1. WAN accelerated backup copy jobs involve both repository servers (source and target) as well as source and target WAN accelerator servers (could be the same as corresponding repositories, if you assign WAN accelerator roles to them). The job is managed by the Veeam B&R server where it is created.

2. Please see system requirements section.

3. Since WAN accelerators are effectively trading disk I/O for WAN bandwidth savings, target WAN accelerator's cache speed is typically the primary bottleneck, not RAM or CPU.

[hollow]

Wow these are some nice links! Thanks alot! Bookmarking it!

[hollow]

Hi Foggy,

Question:

3. Since WAN accelerators are effectively trading disk I/O for WAN bandwidth savings, target WAN accelerator's cache speed is typically the primary bottleneck, not RAM or CPU.

Can I check what is the optimal way to configure offsite backup jobs?
Eg: Create 1 backup-copy job with many VMs? or Create multiple backup-copy jobs (staggered?) with few VMs inside per job?

[foggy]

From performance perspective, there's no difference.