Understanding WAN Accelerator traffic flow

[stevenrodenburg1]

Hello,

I'm trying to get a better understanding of the directions of traffic in a Setup with WAN Accelerators. We have the following setup in the Lab:

1. VM as Veeam Master. Is not used as a proxy. Configured as the Source WAN Accelerator.
2. 2x VM as Proxy using hot-add.
3. A local, fast Windows 2008 R2 Server, serving as the local CIFS based repository and Veeam NFS host.
4. A remote Windows 2008 R2 Server, also serving as repository and configured as the target WAN accelerator, on the other side of a 10Mbit WAN link.

Question: how *exactly* does the traffic flow with a backup copy job, configured to use the WAN Accelerator.

Does it go like this: 3 -> 1 -> 4 ?

Is it better to relocate the source WAN Accelerator? if so, to where ?

[foggy]

Steven, your understanding is correct, data is retrieved from the repository by the transport service installed on it and transferred over the WAN between accelerators. I understand your concern regarding additional hop between repository and source WAN accelerator server, however, considering that typically a primary bottleneck is WAN or a target storage, this should not affect the entire process performance.

[stevenrodenburg1]

Hello,

Foggy wrote "and transferred over the WAN between accelerators."

That was in 2013. We are at v9 now and it does not flow between the two accelerators. It works like i'm describing in this post: veeam-backup-replication-f2/target-wan- ... 35117.html

Which is a problem in our networks design. My question in my new topic is still open and it's best to leave this old topic to rest and continue in the new topic.