Close this search box.

Databases and VMs Only Now Fully Leveraging Disk-based Backup Targets: Interview with Sepaton Director of Product Management Peter Quirk Part I

Ever since using disk as a preferred backup target gained momentum in the late 2000’s, there have been those who opine that disk’s life in this role would be short lived. But those providers who deliver disk-based backup solutions and are betting their future on them see no slowdown in their adoption. In this first interview with Sepaton’s Director Product Management, Peter Quick, we discuss how databases and virtual machines (VMs) are just beginning to take full advantage of the benefits that disk offers as a backup target.

Jerome: Peter, thanks for joining me today. For my readers who may be unfamiliar with you, can you provide a brief background about yourself?

Peter: I’m a veteran of the IT industry. The last five years I have been director of product management at Sepaton. Prior to Sepaton, I did a couple of tours of duty at EMC as well as a small software company in between my stints at EMC. I was also involved in product management and product marketing for a variety of products at Data General. During that time I gained significant exposure to the storage industry to include storage software, graphical interfaces, operating systems and God knows what else.

Jerome: Thanks for that background. Sepaton has been in disk-based backup space for a long time. In that vein, I’d like to get your take on disk’s long term viability as a primary backup target (as a replacement for tape) and how that role played into the significant investment that Sepaton made into its newly announced VirtuoSO platform.

Peter: If you think about backup, there is essentially a disk model and a tape model out there today. Disk and tape technologies serve quite different and complementary roles. Disk is the fastest and most flexible device to land data on. When you add deduplication technology to it, it brings the cost of storage to a competitive level with tape when you take into account various other management costs for tape. For instance, there is more labor involved in managing tape.

But disk does not really have the same retention capabilities as tape. You can store tape for a very long time. We see disk-to-disk backup playing a role in short to medium term retention environments, where a medium retention term might be one to two years.

If you need to retain data for a really long time, seven years or longer, then you are probably going have to include tape in the discussion, just because the average lifetime of a disk these days is around five years. Technology churns are probably even more frequent than that. Tape is going to give you a long term stable storage solution for those very long retention requirements.

But beyond what you are actually storing the data on, you get to the backup model that the backup apps are using, and there are both tape oriented and disk oriented backup applications with many of them supporting both behaviors.

What is interesting is in the very large backup world such as you find with Oracle RMAN and SQL Server, these databases have their own integrated backup capabilities which allows them to write to disk or remote disk, without the involvement of another application which behaves as a media manager.

Prior to the use of disk-based backup for Oracle, etc., people would use NetBackup or Tivoli Storage Manager as a media manager to position tapes and manage the tape catalogs for RMAN or SQL Server. At the moment Sepaton sees a strong desire to separate database backup from general backup. People can avoid licensing costs for the backup software to act as the media manager application if they use the inbuilt backup capabilities of these big database apps.

The second motivation for using a disk as opposed to a tape model is that in virtual environments there are a lot of snapshots involved. It is easy to snapshot to disk but very difficult to do snapshots to tape. It’s just not a natural conversion. When you look at the virtualization of data centers and the increasing role of virtual machines (VMs), clearly a disk model is the way to go.

In summary, disk-based backup provides a more natural operational model for many applications and end-users, the performance is great and can be scaled by adding controllers and spindles. The economics are more than competitive with tape when you add in space-saving technologies like deduplication and data can be replicated to multiple remote sites at speeds limited only by your communications infrastructure.

In part II of this interview series, we discuss Sepaton’s Virtuoso platform and how it was specifically architected to meet these emerging demands of disk-based backup targets.


Click Here to Signup for the DCIG Newsletter!


DCIG Newsletter Signup

Thank you for your interest in DCIG research and analysis.

Please sign up for the free DCIG Newsletter to have new analysis delivered to your inbox each week.