In fact, the latest innovation from us - nobody else has it - is called clustering. It's called clustered data ONTAP as a solution. What it essentially does is this: Look at the traditional way in which storage is procured. You buy a box, and as long as the box allows you to expand, you can expand. And the next time you buy another box, it is treated as a separate entity - that's how most of the deployments happen today. But with clustering, what really happens is, you buy something and then you extend on this using another box; which means all the operating system that runs on all this is one common operating system. It allows you to plug in as much as you want, as and when you want, without creating any bottlenecks for you and without creating any interruptions for the applications. This is groundbreaking, and we worked almost eight to nine years on this before we came out as a technology. It's called Clustered Data ONTAP. It doesn't have a parallel in the industry and it's very similar to what happens in the servers. You can treat this as a kind of virtualization. If you look at the server space - you have a box, you can run 10 virtual machines on this. If you have another box, you can have another 10 virtual machines on this. But as a user, I see it all as a single space. I can see twenty virtual machines which I can run, although physically they are two different boxes. The root of that is in clustering. Because they are able to cluster this, the virtualization is able to rein on the cluster and give you a picture that it is one single physical entity. Very similar to that, is what we have achieved on the storage, which I have said, it's unparalleled in the industry today.
Is it actually using virtualization technology?
Virtualization in storage is slightly different. It comes in different forms. You have different forms of storage, unlike compute. Compute: there is only one CPU, and that CPU needs to be imaged, and you virtualize that image. But in storage, you have different types of storage - tapes, hard disks, Flash, etc. You also have different types of controllers. So virtualization in the traditional sense of the word is to virtualize these different types and provide you with a single view. So for you, it is a memory which is available, but behind the scenes, it will move data from one to another. So that as long as it's hitting a threshold on the Flash, it might move the data to a start-up drive. It's kind of managing your data across multiple types of devices, which is built because of cost efficiencies.
Sign up for Computerworld eNewsletters.