When it comes to data protection integration, Dell EMC and VMware are tightly aligned. Dell EMC data protection platforms have variable-length and source-side deduplication that reduce network bandwidth, improve storage efficiency, and shorten backup times. And if we look at all of the OEMs protecting VMware, the common factor is VADP.
VADP: vSphere Storage API for Data Protection
vSphere Storage API for Data Protection is the API that VMware provides to every data protection vendor to handle image-based backups. That’s right, all OEMs have a level playing field in terms of what they get from VMware to take snapshot-based backups. But wait, if there’s a level playing field, then how can we find differentiators in performance and features? Let’s dive deeper and find out.
This 8-step process makes up every VMware-based image-based backup, so how do vendors attempt to differentiate themselves? The answer is features and transport modes. When the data gets pulled out of VMware, it can either be pulled in full every time or we can do a full and then incremental-like backups using a feature called Changed Block Tracking (CBT).
CBT provides the ability for vendors to do a basic level of reduction of the data at the source by only sending the blocks that VMware marks as changed since the last backup. One way that Dell EMC and a few other vendors (Commvault, Veeam) differentiate themselves from the rest of the pack is that they can use CBT on restore, not just on backup. As we move onto the transport modes though, you’ll see that there are even better ways to handle data reduction.
VMware transport modes
There are three transport modes that are provided by VMware to move backup data out of vSphere and into a vendor’s backup application. They are SCSI Hot-Add, SAN Transport and NBD.
Dell EMC uses SCSI Hot-Add as the default transport mode in their vProxy-based backups, with the ability to fall back to NBD to get a good backup if needed. SCSI Hot-Add allows Dell EMC to put their source-side deduplication engine into the proxy and perform the data reduction globally before the backup data is sent to protection storage.
The benefit here is that the least amount of data possible is being transferred over the wire, and the backup times significantly improve. This also means that the snapshot lives for a very short amount of time, leading to less disk utilization used by the delta disks during the backup process.
In the new PowerProtect X400 appliances, Dell has embedded an internal vProxy that uses NBD in order to get VMware backups running quickly. However, for any sort of scale, I’d recommend leveraging their vProxies in order to have access to SCSI Hot-Add.
Though this method has become less popular the last few years, some backup vendors leverage SAN transport mode, which allows for a LAN-free backup. With this method, the snapshot data is moved over Fibre Channel from the SAN to a physical proxy server.
The largest limiting factor to SAN transport is that it requires a physical endpoint into the storage and with newer HCI technologies, specifically vSAN, it’s not supported at all.
For the last transport mode, let’s talk about NBD, which stands for network block device. You’ll see this a lot, as it’s the standard and only transport method for some OEMs, while most have it at least as an option.
NBD requires that all data be moved over the wire, with any reductions only coming via CBT. This means that any data reduction (compression, deduplication, XO-Differencing) is all done target-side, whether it’s inline or post-process. In addition, all NBD traffic is sent through the VMkernel NICs, which are also responsible for management traffic, etc.
Going through this material, I’ve mentioned both hardware and software proxies for backing up data as well as the different transport methods made available by VMware, with a particular focus on the VProxy SCSI Hot-Add approach from Dell EMC.
All OEMs use proxies to perform image level backups, but as you’ve seen above, there are a few different options available depending on the backup application. If you would like to dive deeper on this or any other data protection topics, please reach out.
We'll be testing this tech out along with all of the other systems we have available in the ATC. Make sure to follow our Secondary Storage topic to stay up to date with the latest from our lab experts!