Since it's a direct attached array, the "network" bottleneck actually is just the source and target datamover talking inside the same Veeam server.
Seems like source and target are more probable bottleneck, but are these jobs running at the same time? Do you have a job running alone so to see the pure performance without the concurrency in place?
EMEA Cloud Architect @ Veeam Software
Veeam VMCE #1