Backup, backup, backup: it’s what everyone’s talking about these days. You have to save all those files, for one reason or the other, and probably because someone is telling you that you have to save those files.
I recently sat down with Joe Disher from Drobo and Storage Switzerland Lead Analyst George Crump to talk about backup, and to review our webinar “The Four New Requirements for Disk Backup and the Virtual Environment” now available on demand.
Charlie: George, we always heard that disk backup solved a lot of problems for people needing to protect their data. But, why do we have four new requirements for backup now?
George: Well Charlie, one of the things that disk backup did was really making the whole backup process faster. It enabled capabilities like faster restores, and faster backups. As we move into the virtual world with VMware and things like software products – like those from Veeam and PHD Virtual and others – came out with features like change block tracking and recovery-in-place further reducing backup times and further reducing recovery times. However, those solutions changed what we expect from backup devices.
Now we need them to be able to handle a random I/O stream, and we need them to be able to provide production class performance to features like recovery-in-place. What Storage Switzerland has done is identify four new requirements for disk backup that many disk backup solutions lack today.
Watch the webinar “The Four New Requirements for Disk Backup and the Virtual Environment” available on-demand.
Charlie: Joe, how does Drobo fill these new requirements for disk backup and virtual environments?
Joe: By being able to provide a device that operates at the speed required, not only for backup but for recovery, while at the same time not sacrificing capacity. Most importantly, being able to provide those things seamlessly without jumping through a lot of hoops.
It’s making sure that you can provide the product at a cost that doesn’t break the bank. Also, being able to utilize off the shelf drives and leveraging existing software packages, such as dedupe, which can further keep the costs down.
Speed is also important, but being able to do it affordably. It’s really important, and we’ll cover that on the webinar.
Charlie: George, we’ve mentioned two of these requirements – faster and cheaper. What are the other two?
George: First, we need faster systems that are enhanced with solid state disk so that they can receive the small write I/O of change block tracking and we need that SSD to be able to mount VMs that have been recovered in place.
Second, we need to drive down price. However the problem is that, like change block tracking in particular, it is getting conflict with some of the other technologies we’ve used to drive down price. So, the only thing left is to reduce the raw price of the device itself. We can’t trick software, we have to actually enable the use of off-the-shelf drives.
The third requirement is being a production device for a period of time. All of a sudden these devices have to be very good at mounting and hosting production data. You don’t want to have an enterprise class system because that means enterprise class prices. Yet, you need to have just enough redundancy that you can run the virtual machine and feel comfortable with it, that it’s bringing your production system back online.
Finally, it’s got to be scalable. We’ve got to be able to grow the system in a very cost effective way. Joe, I think that’s something that you guys do quite well, is fill all four of those niches, true?
Joe: Yes, absolutely. I think that the key to all of this is faster recovery of data and virtual machines in place, having the ability to do that faster in a cost effective way, and also present the production. Most importantly, the ability to scale that storage as needed is also critical. One of the things we’ve known for a long time related to unstructured data, like files, is the amount of data being loaded up is enormous. It happens all the time. Even in a virtual machine environment the amount of data being loaded up on these virtual machines is growing exponentially over time.
We offer scalable back end backup infrastructures that can scale with you simply by sliding out smaller drives and sliding in bigger drives as your environment changes, without having any downtime. That’s a critically important piece of the puzzle. As we all know with backup windows basically growing to zero. You don’t have time for your system to shut down for a weekend while you’re doing that. So I think that’s certainly another key element, the scalability.
Charlie: Joe, do you see customers starting to use your systems for backup then start using them for production?
Joe: I see it goes both ways, but I think more often than not folks lean towards whatever they happen to be working on. If the problem they’re trying to solve is a backup problem, then we can solve that problem upfront and fill in that niche. It’s just a matter of being comfortable with the product and then realizing the capabilities.
We see them start to shift over their resources for production data, and vise versa when they realize we have the ability to grow capacity the way that we do. They’ll start to recognize the advantage of utilizing that extra capacity the way that they’re familiar with backup, if they were using it for production data. So it does go both ways, but I certainly see it being used more often for backup today.
Charlie: So recovery-in-place is not bad?
George: No, I don’t think so. As you know I do those data protection workshops and we talk a lot about recovery-in-place. I’m generally advising people to be aware of these issues. Just like in any area of IT there’s always a solution, you just have to know where it is. Drobo is a really good example of that. This is a device that really meets those requirements very comfortably.
Editor’s Note: The webinar is available right now. Click here to get it any time on-demand.