Understanding VEEAM Network traffic

[HarleyC]

Hi Everyone

We are currently using VEEAM B&R 7 R2 Enterprise Plus, and have been using it for a while now with great success

I'm trying to do a bit more work on our infrastructure and am looking at working tape backups into the VEEAM solution

I want to make sure that I understand the network communications between the different tiers of our current backup solution so that I know where I can improve things for the future

Our current configuration is as follows

3 VMware hosts with a total of around 70 VMs
All of the VMs are running from a DELL EqualLogic SAN

VEEAM B&R is running as a VM within the setup above, and all of the SAN luns are shared through to the VM directly within Windows so that it can do SAN based backups
VEEAM Backups are stored on a separate Windows 2012 NAS server over a 1Gb LAN connection

Each night we have the following jobs running
- (1) nightly backup of all VMS (forever incremental with synthetic fulls once per week) ... 60 days retention
- (2) backup copy job that creates a backup set of a single full from the above each day
- (3) backup copy job that creates a retention set (yearly, monthly, weekly)
- (4) replication job that replicates vms to another server
- (5) tape copy job that copies the repository created from (2) to tape each night (STILL TO BE ADDED ... at the moment I'd have to connect tape drive to NAS and connect to VEEAM server through iSCSI)

When running the backup copy jobs, does all of the backup data flow from the NAS, to the VEEAM server, and then back to the NAS again? Or do the VEEAM components installed on the NAS deal with this locally?

I'm trying to decide if there would be much benefit in moving the VEEAM backup server components entirely over to the NAS server and running everything locally from there rather than having the current split of VEEAM/Storage. The fact that the tape drive is connected to the NAS points to just installing everything on the NAS but I'm wondering if the other jobs would benefit also

[foggy]

Harley, in direct backup copy job mode data is transferred directly between agents installed on the source and target backup repositories.

[HarleyC]

ok, so does that mean that a backup copy job that is copying from a repository on a NAS server to a different repository on the same NAS server keeps all of the data locally on that server and doesn't need to send it all over the network to the main VEEAM server?

[foggy]

Correct.

[HarleyC]

ok, great

Thanks for the quick replies

[Gostev]

Just to avoid further confusion for future readers, this actually not in case with what is normally called NAS. Since typical NAS boxes cannot have 3rd party agents installed, or execute any data management commands for that matter, the traffic will go to the server running the data mover, then back to NAS.

What makes it possibly in this case, is that what you call NAS is not actually NAS, but rather a standard Windows Server with local storage. As such, this server can run our data mover component, which will perform all data processing right on the box.

But, if for example you setup your backup repository using the "shared folder" option (as opposed to "Windows server" option), you will end up in "true" NAS situation with data traveling across the network, just as I described above.