multithreaded data transfer

[MAA]

Hello,

https://helpcenter.veeam.com/docs/backu ... tml?ver=95

Does this feature (multiple TCP/IP connections) work only for backup job, or for all jobs (include restore job guest files)?

[Andreas Neufert]

Hi, this is for inter Veeam communication only. For example between Proxy and Repository.

But our Proxy read multiple VM disks in parallel and use asynchronous operations (multiple requests are sent instead of one) with Virtual Appliance mode, Backup from Storage Snapshots and DirectNFS.

[MAA]

It means the restore job (guest files) is not using a proxy, and is a single-threaded?

[Andreas Neufert]

The connection between repository (or gateway server) and Mount Server uses this setting then.

Basically this setting affect the inter Veeam communication (only).
Usually a change has in normal operation only an affect if Bottleneck analyse shows „network“ as bottleneck.

If you transfer backups over WAN links without our WAN accelerator you can potentially increase throughput with it when WAN link latency is high.

If you have All Flash source and target storage and try to optimize transport for 3GB/s or higher, this setting is sometimes helpful to set higher.

[MAA]

I have suggestion for future B&R version - multithreaded restore job.

current version:
[Repository]---<single-thread>--->(WAN link)---<single thread>--->[VM file server]

My suggestion for future version:
[Repository]---<multi-thread>--->(dual WAN links)---<multi-thread>--->[Proxy]--->[VM file server]