Rachel Dines, an analyst at Forrester Research, discusses the pros and cons of disk-to-disk-to-cloud backup and hybrid cloud backup approaches in this Q&A. Read her answers below, or download the MP3 to help get a better understanding of hybrid cloud backup.
Listen to the hybrid cloud backup Q&A
>> Can you outline what a typical disk-to-disk-to-cloud backup
scenario might look like?
>> Have you seen much adoption of disk-to-disk-to-cloud backup yet?
>> What are the benefits of this hybrid cloud backup approach? Why move away from tape?
>> What are the drawbacks of disk-to-disk-to-cloud?
>> What vendors are currently offering disk-to-disk-to-cloud backup?
There are actually two ways that I see disk-to-disk-to-cloud deployed. The first way is a traditional on-premise backup deployment with whatever enterprise or small- to medium-sized business (SMB) backup solution you have, whether it's with CA, CommVault, IBM Tivoli Storage Manager (TSM) or Symantec Corp. You would back up your data locally and then take your local backup to disk, and replicate that to the cloud as if it were any tertiary medium like tape. So that's very well integrated with what you have today. It's not really a huge departure or radically different from what we're used to with disk-to-disk-to-tape, which is a very common setup.
Another disk-to-disk-to-cloud deployment model that I do see is a managed appliance from the provider where they put an appliance onsite; in this case it would be more like a backup-as-a-service, except you would still be backing up locally to an appliance and then replicating to a cloud. The main difference is that you're not managing the software at all. You're not managing the appliance. That's all done by the provider.
I've seen a fair amount of adoption. People seem to be a lot more comfortable with disk-to-disk-to-cloud rather than directly going to the cloud or backup-as-a-service. Our most recent data from over the summer found that 23% of companies said that they were using some form of disk-to-disk-to-cloud. And that's actually quite a bit. We also had a lot of people saying they were interested in it and investigating it. That's actually a lot further along than just pure backup-as-a-service or just disk to cloud.
The type of disk-to-disk-to-cloud backup companies are deploying depends on their size. Larger companies seem to be taking the model that I first described, which was to use their traditional on-premise backup and just replicate it to the cloud. Smaller companies tend to use the managed appliance from a supplier. But either way, the disk-to-disk-to-cloud is the much more dominant model in online backup.
There are a few reasons to use cloud over tape. The main benefit of going disk-to-disk-to-cloud is that you're automatically getting your data offsite. This is great for resiliency for disaster recovery and having your data go immediately off-site. On the other hand, if you went with disk-to-disk-to-tape, you'd have to take the tapes, encrypt them, and ship them off-site and deal with all of the physical movement.
One of the main reasons I see people considering a hybrid cloud backup service is because of the easy and quick digital transportation. There's also the fact that you don't have to handle the physical tapes. Getting tapes in and out of a library is time-consuming. Those are the main reasons that people tend to go that route, although there's definitely downsides to going with disk-to-disk-to-cloud rather than disk-to-disk-to-tape.
The first drawback is price. It's really hard to beat the price of tape. There's no storage medium that's cheaper than tape. We're seeing cloud storage get down there, but tape is still the cheapest. There are also all of the risks that you have with going with any service provider. That's not to say that it is extremely risky, but I find that a lot of people get nervous about putting their data in the cloud, and there's fair reason for that. There's the multitenancy aspect, and they're contracting with a third party, and so there's a lot of due diligence that has to be done around that. They must ask questions such as: What are their security models? How are they protecting your data? How are they storing your data? And are they using your data for something they shouldn't? So there's definitely an added risk in going to the cloud model, but on the other hand, there's the argument that you can lose tapes. You have to worry about encrypting tapes. And sometimes you can lose tapes when moving them off-site. So I think there's definitely more of a perceived risk of going to the cloud, but I'm not necessarily sure that I believe that that is actually a true risk.
In the traditional on-premise backup provider space, the main vendors I see going disk-to-disk-to-cloud in enterprises are CommVault, EMC NetWorker, IBM TSM and Symantec NetBackup. These are really the main players, and there are a lot of them that can offer cloud as an eventual target for data. In the SMB and midmarket space, Acronis Inc., Asigra Inc., BakBone Software (recently acquired by Quest Software), i365 (a Seagate company), and Symantec with Backup Exec can also go to the cloud. So it's becoming a pretty common feature in on-premise backup solutions that they can use the cloud as a target. And it's not necessarily going to be any cloud -- they usually have partnerships with different cloud providers that you can work with, or else maybe you can back up only to that company's proprietary cloud.
From the appliance perspective, providers that offer disk-to-disk-to-cloud offer it as a managed appliance that sits on-premise that you backup locally to and then replicate off-site. There are hundreds of vendors that provide this option, but the main player that I see in that space would be i365, and Barracuda Networks (they acquired this technology through their acquisition of BitLeap a few years ago). Iron Mountain Inc. can do disk-to-disk-to-cloud as well, although they're traditionally more situated in the backup-as-a-service market. These are the main players, but there are hundreds of them, and lots of resellers as well.
This was first published in January 2011