The first episode of Raising the Floor, the enterprise datacenter podcast from Foskett Services, features thought leaders from two key cloud storage-enabling companies:
We continue our conversation talking about some of the advanced features of cloud storage gateways: Data protection with snapshots, automatic provisioning, and easier storage management.
Stephen: So you are tackling the file related issues and Josh, I understand that Cirtas is tackling the issues related to block storage. Now, I think that’s actually a different kind of jump than file systems.
Josh: There is some things that are different and there’s a lot of things that are the same. So, in our case, what we are doing is presenting a standard iSCSI target so that you access are going to revert iSCSI block protocols. And the way you interact with the system would be very much familiar to you from using any other iSCSI storage array. So, you would go in provision volumes, everything is still provision. So, you can expose to your observers whatever capacity you would like without that happening any corresponding charge in the cloud other than what you are actually writing in per data.
So, as an example, I can give a server a five terabyte volume to use very easily and from the operating systems perspective on that server, he has five terabytes and disk space but you have not been charge to send by your cloud provider by provisioning that volume. It’s only going to be based on what you actually write into it.
And then again, as Andres mentioned; and this is very important in this world, you want to look at only the deltas of information coming in as far as what’s going up to the cloud. So this system needs to be able to de-duplicate and compress information as well as to version it.
And in our case we, like Nasuni does; take snapshots of data on each volume at user defined intervals. So this gives you logical data protections layer that overlays on top of the physical data protection you’re getting from the cloud providers.
Those two things together are why you don’t need to do separate backups when you’re using these systems as primary data stores. The protection that is inherent in backups is just there for you automatically without having to run the backup process.
Stephen: Yeah because you’re using snapshot technology as an alternative data protection method.
Josh: That’s exactly right. So the two elements of data protection are going to come down to physical data protection, in other words, what do I do if the driver or site fails? That’s what the cloud providers are doing. And then there is the logical data protection, which is what do I do if somebody deletes a file they care about? Or we get a virus and it causes a problem, database gets corrupted? In those kinds of events you’re going to do a logical rollback even though the underlying storage system hasn’t had any physical problems to it.
Stephen: So that is where snapshotting technology is critical?
Andres: Stephen, one thing I want to add to Josh’s comment. So when he mentioned the ability to provision say a five terabyte volume, but your paying for as much as you get in the cloud. One of the really interesting things about that is that Josh is essentially describing thin provisioning in the classic sense. But unlike doing thin provisioning and this is the cloud is automatically provisioned. So he is essentially provisioning the volume. But then the volume is growing gradually, smoothly without any administrative interference. Without you having to worry about it; the volume is filling in its data as it goes. And that is one of the things we talked about in the beginning.
The cloud really allows you to build a different type of storage system, because automatic provisioning is really thin provisioning should be but isn’t. When you’re talking about physical linked arrays that are limited by actual hard drives, you know, running, spindles running in your datacenter.
Stephen: Yeah, that’s one of the items that I am trying to basically get across to the world. Cloud storage is the next generation of thin provisioning. It’s this automatic provisioning, this idea that as I need storage I get it and as I stop needing it I don’t have to pay for it anymore. And that is really revolutionary; I mean that’s a really totally new thing. Does that give you the kind of economics that people want from storage though? I mean, is de-duplication and automatic provisioning, is that enough to justify the cost of cloud storage. When you can go to a store and spend a hundred dollars and get a terabyte of disk storage?
Andres: So you’re never going to be able to compete with Best Buy drives. They’re cheaper; the cloud is more expensive than that. If you look at the Amazon costs by themselves, you’re going to be looking at about five dollars per gigabyte over a three-year period. That is very, very close to the cost of primary storage array technology. If you were talking about a left-hand or an ecological array, you would be talking in that range. Now what is interesting about the cloud is even if you are paying about the same amount for the primary piece, the load that you are taking off the protection environment. Everything that you have behind that to protect that storage. Whether it’s a second site, whether its your backup server and targeting say, some data domain or some data de-duplication box in the backup stream. That what gets above simpler because now you’re relying on the cloud for doing all of that work.
So I think when looking at the cloud for cost, it’s very important to not just stop at the hard drive or primary cost level and really try to look at your cost as it ripples through your environment for protection.
Stephen: Because there’s even more than data protection as well. You have to think about storage management and planning and migration forward. That’s another thing I think people often overlook is the fact that you can freely migrate forward using cloud storage because there is no migration. It just exists and when things are updated, they’re updated and you just never even know it. You don’t even notice that things are happening.
Josh: What Nasuni and Cirtas are addressing are all of those things. It’s not just the capacity in giving you inexpensive capacity. But it’s making the whole cost of your data life cycle with really solid data protection and really solid disaster recovery. Something completely viable that you can do with the cloud. So, in our case, you’re going to be able to store data. That data is going to be protected in the same way as if it had been backed up but you’re not going to have the costs of developing those backup systems.
You’re also not going to have the costs of building out and maintaining a disaster recovery site in the traditional sense where you would have synchronous or asynchronous replication to a remote array.
Once your data is in the cloud, it’s accessible from anywhere on the Internet. So, if there is some kind of disaster and you need to recover straight out of the cloud, that’s perfectly doable from any location. The need to have very expensive disaster recovery processes is changed dramatically by leveraging cloud technology.