Backup repositories, i need an advice

[TrK]

My customer have an equipment that can be used as backup storage:
Оne powerful enough and still good server with E5-2420 CPU, 32GB RAM and 5.45TB storage (4x2TB SATA HDD in RAID5). OS - ESXi6.
One powerful enough and still good server with E5645, 96GB RAM and 5.45TB storage (4x2TB SATA HDD in RAID5). OS - ESXi6.
One pretty old and slow server with entry level AMD CPU, 8GB RAM and 5.45TB storage (4x2TB SATA HDD in RAID5). OS - Windows 2012R2 (can be converted to Nexenta or something if needed).
E5-2420 and old AMD servers can be placed in any of two sites (main and backup), E5645 is planned to use only on backup site. Site link is 1Gb with low latency (not WAN). Backup site is used for backup storage and some non-critical task.
Veeam Backup is planned to install as VM on some backup site (main or backup).
Full backup is 3.7TB in size, increments - 200-300GB. Customer wants to keep minimum of 14 increments (30 is better) on backup site and last full on main.

I can only propose to use Scale-out repos, but cannot understand how to distribute equipment and roles between servers. Can someone give me advice on it?

[foggy]

Hi Dmitry, just to get a better idea, do you mean you can re-purpose the two ESXi servers to serve as backup repositories for Veeam B&R, along with an old server? Otherwise I don't get the idea with scale-out repository in this setup.

[TrK]

Yes. Whole E5-2420 can be used as backup repository. Аnd E5645 whole RAID5 LUN (CPU and RAM can be used partially).
Oh, i forgot - customer need copy of full backup on second site (for DR) or ... replica of all VMs?

[foggy]

Then I suggest to introduce E5-2420 and the old AMD server as extents for a scale-out repository and use E5645 either as backup copy target (to keep the full at DR) or for replica VMs. This tool for space estimation might be handy at planning stage.