Backup 200 GB over WAN ?

[liviu.tutuianu]

Dear all,

One of our clients wants to backup their critical data over the WAN to HO.

Infrastructure:
33 Branches with 1 ESX Host
1 Head Office (HO) with a new dedicated VBR - physical server (HP Proliant DL380 G8; CPU: 2 * 8 Core; RAM: 64 GB, HDD: 2 x 300 GB (for Metadata) + 1 DataDomain storage of 18.3 TB
Network Bandwidth : 2 Mb (256 KB)

Data to be transferred from each branch : Initial backup: 200 GB .
Incremental after that around 1.5 GB per branch

Also, we may implement WAAS (WAN Accelerator from Cisco) to reduce the time transfer between branches and HO. We don't know exactly how much this will help...

Data that will be transferred: as this is a fileserver, only regular documents will be transferred (.xls, .doc, .pdf, .jpeg . etc) , no databases or e-mails...

And now, the questions:
1. Using VEEAM Backup Jobs will compress in any way the data that is being transferred over the network, or the files will be sent uncompressed. I am asking this because if we will use WAAS, than WAAS will also try to compress the data before sending it through WAN, and I don't know exaclty how much will help us WAAS in case that VEEAM also compress the data.

2. For the same servers, we will have actually 3 backup jobs (each one for Daily, Monthly and Yearly Backup), each one with different retention periods. Is there any way by which after the first full backup, to use that fil for the Monthly and Yearly Backups? Basically, the data will be quite the same, but in this way, we could reduce very much the network traffic
for the initial run of these 3 jobs.

Theoretically, the biggest problem will be the first full backup. After that, the VEEAM solution should work fine, right?

Thank you in advance,
Liviu

[foggy]

0. First of all, the metadata should be placed on a repository that is close to the source proxy, as explained at the corresponding step of the job wizard.

1. Transferred data is both compressed and deduped (plus user guide, p.31). Also, you can search this forum for multiple existing discussions on using WAN accelerators.

2. You need to seed the backup for your remote jobs.

[jpeake]

If the VM's at each branch are Win servers, and you are just needing to back up file shares, maybe DFS-R back to the HO, then Veeam doesn't have to deal with all the branches. You can just back it all up at the HO. DFS-R seems to work pretty well over slow links.

Just a though

[dellock6]

Yep, also because the initial assumption is wrong: even if the server is "only" a fileserver, Veeam does not works at the file level, but at the VM level, so you also need to take into account the blocks changed in other parts of the VM, like OS files for example. As usual, a good test is to run local backups and see the size of the incremental backup.
About wan accelerator, we worked with other vendors (riverbed and bluecoat) and we see additional gain in using them, they do apply caching and not only data compression, so they can help. I cannot tell you how much however, never used Cisco ones.

Luca.