A little help with Best practices?

[szwicker]

So, I have network with about 35 VMs spread across 4 hosts in a DRS cluster with 2 Dell MD SANs.
I have several standalone boxes with decent hd space & processing power.
As of now they have just been running backups & rep jobs using the basic network proxy.
I want to clean things up.
Should I convert Veeam to a VM on the actual production network? Keep it as is but add additional NICs to the SAN switch and pull backups directly from the SAN instead?
Thoughts?

[Vitaliy S.]

Hi Seth,

What is your current processing rate? What does bottleneck stats show for your backup jobs?

Thanks!

[szwicker]

Bottleneck is source. Speed is around 50/ MB/s ranging between 40 - 65 I've seen.

[szwicker]

A little more info:
One of the 3 BDRs has a about 30TB of space. It's running vSphere 6.0 with a Win2012 VM that's running Veeam 8.
The VM has a proxy (network) and repository (15TB VMDK) for onsite backups It's the master server. The remaining 15TB is used for replication jobs.

I have 2 dual 10G NICs ordered. I intended to plug those directly into the 10G SAN switch to improve performance.
How do I make optimal use of that NIC with the server intended to receive the replicas?

In terms of the backups, is it more trouble than it's worth for me to have multiple standalone servers with their own datastores/proxies?
Should I pull the drives (leaving the bare minimum) and just use the servers (decent processing power) as proxies with one acting as central storage for backup?s