Q&A: EMC's Gahagan on taking a virtualization ride

ANAHEIM, Calif. -- Who knew storage management could have anything in common with a Disneyland ride?

It may not be the first parallel you'd draw, but in a keynote address Tuesday at Storage World Conference 2003, EMC Corp.'s senior vice president of storage infrastructure software, Chris Gahagan, explained how a good virtualization strategy operates in a fashion similar to the virtual reality ride at Disneyland. The theme park ride, which tricks riders into thinking they're hang gliding over California, is able to make the illusion real by providing sensory clues to passengers, while keeping the mechanics of the ride hidden. A good virtualization tool does the same thing to your storage, though perhaps somewhat in reverse: It sets up a matrix in which the data thinks it's traveling on a familiar path, while enabling the operator change the ride at will.

In an interview after his presentation, Gahagan elaborated on his vision for virtualization, which on the EMC road map, is known as the AutoIS strategy.

What do you see as the next step in automation? What is the next thing customers want?
[There are two areas.] From where I sit, the next thing they're going to want to do is control allocation. They want to be able to treat their storage independent of their applications. Maybe a better word is flexibility. How can they have flexible control over their storage environment? That's really what our customers are asking us for, is that flexibility to take their resources, their hardware, on the virtualization ride and reprogram it with different rules while the application is running and [have] it become much more flexible. What's the other area?
The other big area we're working on is something called information life cycle management, which is managing the data that you have from the time it's created until the time it dies.

Here's an example. If you're a cell phone company, the data that you capture about cell phone calls is extremely valuable for you -- for some period of time. This is typically until you bill your customer and then for some period of time going forward, until the bill resolution happens. [But] that data is still valuable to you, for example, if you want to look at what cell zones get the most activity or if you need to put more cell towers in.

But the value's not the same to you as if you hadn't collected the money yet, right? You still want to keep the data around, but how do you keep it around for less money? How do you put it on storage that doesn't perform as well? And when do you need to retire it?

Information life cycle management is technology that will monitor the data from when it's created, knowing when it should move from this class of storage to another class of storage and when it should be deleted. So it's all the aspects of, I'll say automating, but not from a technology perspective -- from a business set of rules. Really, that's what customers want to say: 'This data has this value for me at this point in time. After a certain point of time, it has less value. At this point in time, I want it to die.' Can you offer any real-world examples?
Companies that harness more information ultimately win. Here's an example: Wal-Mart and Kmart. [Wal-Mart] does tremendous amounts of data mining. Every single item you buy gets loaded into a data warehouse. They look at buying patterns: what you buy, when you buy, what time of day you buy.

And then you take Kmart. They don't do any of that. Kmart and Wal-Mart sell the same products, from the same vendors, for roughly the same price. How Wal-Mart is winning is on margin. They know how many items to have in the store. They use this tremendous power of information.

That's this whole thing about information life cycle management. How do we take the information that companies gather -- that they have to gather to do their business -- and give it to them to use in other ways they haven't thought of? It's real exciting stuff. People find storage boring. It's totally sexy. EMC has made a number of acquisitions recently to augment its software development efforts. Do you see more acquisitions on the horizon?
We're looking at everything [including purchasing companies, as well as in-house development]. I think the important thing is, what's the balance? Are we doing a good job of acquisitions where it makes sense? Are we doing a good job of building applications, or software, from the ground up? Some companies only do one -- NIH, not invented here. I think we're doing a good job of balancing, which is the key. Do you think that there's a point at which vendors can give customers too much automation? If so, how do you track where that point is?
We actually do see that. We have a product; it's part of ControlCenter and it's called [Automated Resource Manager]. You can actually go to the product and say, 'On this server, I need another 10 gigabytes or 100 gigabytes' or whatever, and it will do everything you need, from talking to the array, to loading it into the SAN (storage area network), to configuring it on your server, without you doing anything other than pressing a button.

But testers said, 'That's not how our business works. We have something called change management. We want to automate, but we want points where you stop, come to the user and say, "I'm going to do this."' So is there too much automation? I think we can do too much without user involvement. The user wants to be in the decision-making process.

Humans by nature are a little bit fearful of giving up control. But I think that, over time, once you prove that it's doable, take risk and cost out of it, they'll come to expect it. People complain that efforts to achieve virtualization seem to go on forever.
People say, 'When is storage management going to be done?' It's going to be finished as soon as the price of PCs stops dropping. It's never. Every technology evolves. What we can do today is a lot better than what we could do, really, a week ago. Are we done? No. What we want to do is build on what we've done. We've solved some problems. Now we're going to solve some new problems.

 Q&A: EMC's Gahagan focuses on storage management

SD2002: EMC offers tool for automating storage provisioning

Users on virtualization: Hold on, I'm not ready yet

Dig deeper on Virtualization Strategy

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

SearchSolidStateStorage

SearchCloudStorage

SearchDisasterRecovery

SearchDataBackup

SearchStorage

SearchServerVirtualization

SearchVirtualDesktop

Close