Veeam presented a couple of new features for their Backup & Recovery suite at Tech Field Day 20. Three of them stand out for me: the Scale-out Backup Repository, Enhanced NAS backup, and Continuous Data Protection. The first one is a logical container to bundle other backup repositories in. The second because backing up large NAS filesystems is hard, especially if you want to maintain a low RPO. And talking about RPO’s, there’s Continuous Data Protection (CDP) which protects your volume or VM with an RPO of seconds. Let’s explore what’s changing and when we can expect to see these features in the field.
Veeam Scale-Out Backup Repository (SOBR)
A Veeam Scale-Out Backup Repository is a logical entity that groups several backup repositories. Backup repositories (or extents) store the actual backup data, and can be of a couple types:
- Direct Attached Storage (DAS)
- Network Attached Storage (NAS)
- Deduplicating Appliances (like a Data Domain)
- Object stores (S3)
- Cloud Connect
During configuration, your backup jobs would all target the Scale-Out Backup repository. The Veeam B&R server would then use policies to determine where backups should actually go to inside this SOBR. This could result in a policy that, for example, pushes full backups to a deduplicating appliance because they get the best storage efficiency. Incremental backups could go to direct attached or network attached storage, because those could be either faster or cheaper. Using Veeam ONE, you could keep track of the individual extents and determine whether jobs are performing better on one type of extent or another. You could then use this information to optimize placement of jobs.
Cloud Tiering
The above types of extents or backup repositories all fall in the Performance Tier of the Scale-Out Backup Repository. There is also a Cloud Tier or Capacity Tier that extends those backup repositories transparently and intelligently to (cloud) object storage. It offers two modes: Move or Copy mode.
- Move mode leaves the metadata in the performance tier, and moves the backup data (in 1MB, source-deduped blocks) to the Cloud/Capacity Tier.
- New in Veeam V10 is Cloud Copy mode, which copies the backup data and metadata to the cloud immediately after the job finishes.
What that enables you to do is to either offload data to the (hopefully cheaper) cloud in Cloud Move mode, or achieve backup data resiliency with the Cloud Copy mode. Anthony Spiteri dives a bit deeper into this cloud tiering, which you can see in the 3nd video over here.
Veeam Enhanced NAS Backup
New in Veeam Availability Suite v10 is Enhanced NAS backup. You want to be able to backup all NAS systems (whether a black box like an Isilon, or a Linux/Windows file server) in a similar manner. There’s also a scale challenge: NAS systems have a tendency to get big, with petabytes of data.
The Veeam Enhanced NAS Backup simplifies backing up NAS systems by offering a single intuitive wizard to back up any type of NAS system. This could be a SMB or NFS share, or a Windows or Linux file server. And it would back it up quickly. The ‘quickly’ is due to parallel processing (read: multiple threads), and more importantly, by using file and folder hashes to determine which files and folders haven’t changed and can be skipped.
Restoring data can be done either as a whole share at once, or with individual files and folders. You could make the latter option available to your users, so they can determine for themselves which old version of files they want.
Continuous Data Protection (CDP)
Lastly, a feature that we’ll see in the first half of 2020 is Veeam Continuous Data Protection (CDP). Think of this as a VCR type backup for virtual machines that continuously records changes. CDP gives you an incredibly small backup RPO (or window of data loss), measured in seconds. This is a vast difference compared to traditional, on-a-schedule backups that run once every x hours.
CDP enables you to locate the ideal restore point of a virtual machine with very fine granularity. And just like an oldskool VCR tape, a bigger journal would enable you could go further back in time.
Using VMware tags, Veeam would automagically add newly created VMs to the CDP backup job and protect them.
Licensing wise, Veeam is still discussing how this feature will be sold to customers. It will likely be a feature license in the Enterprise Plus suite. It could also become a volume license where you buy a license to protect a certain number of VMs. With release in 2020, Veeam still have some time to choose a route.
My thoughts on the Veeam announcements at TFD20
First of all, the Scale-Out Backup Repository. Anything that makes it simpler for an admin to determine data placement gets a thumb up from me. The Veeam SOBR approach makes it intuitive to automatically assign backup repositories to a certain type of backup and/or backup job. You, the admin, determine the rules. The policy engine then does the rest, 24/7. The Scale Out Backup repository is also a linking pin in the tiering of backup data to the cloud. This could help large backup systems to maintain low (backup) storage costs. Or it could help small Veeam customers to achieve backup redundancy by copying backup data to the cloud instead of an expensive secondary off-site backup repository.
The enhanced NAS backup brings some consistency to making NAS backups. Regardless of the source of the data, Veeam gives the backup operator a consistent backup and restore GUI. During the presentation I missed the fact that Veeam uses both a file and folder hash to determine if data has changed. I thought it only used a file hash, and feared a heavy constant read workload on the NAS if Veeam is constantly reading and hashing data. The folder hash helps reduce that read workload a bit, especially on large, static datasets such as PACS medical imagery. And there are several options in the GUI to throttle the backup process.
Still, I am not entirely convinced that this will be suitable for true multi-PB file systems though. Scale-out enterprise NAS systems have some very robust replication and snapshot technology inside. Multi-PB systems are typically replicated off-site anyway, with some snapshots on the replicated data set to go back in time. This typically offers plenty of protection against either physical disasters and/or software/malicious disasters. And very importantly, that approach also ensures a short RTO.
Then again; Veeam Enhanced Backup could offer a good 3rd backup copy of data for those that need airgapped backups on a different technology stack. And it definitely is a good solution if you have a large number of multi-vendor (smaller) NAS shares/appliances in your company.
Continuous Data Protection is always interesting since it offers a low RPO backup that doesn’t hammer your underlying storage & compute systems at fixed times each day. Instead, it backs up 24/7 in parallel to your existing workload. One of the gripes with another CDP solution is that it is a pain to add newly created VMs to protection groups. Things like VMware tags make this much easier, and Veeam fortunately thinks about these things. The product isn’t available just yet though, so I’m hoping to play with it in the future and get a real sense of what it’s worth.
Check out the videos on the Veeam TFD20 page.
Disclaimer: I wouldn’t have been able to attend Tech Field Day 20 without GestaltIT picking up the tab for the flights, hotel and various other expenses like food. I was however not compensated for my time. There is also no requirement to blog or tweet about any of the presentations. Everything I post is of my own accord and because I like what I see and hear.
