From boring to personal: Challenges of Data Management

Zafar Anjum | June 25, 2012

The world of data management has become much more challenging in the past two decades, but in terms of NetApp’s business, that’s a good thing because it gives plenty of room to innovate, says Dave Hitz, executive vice president and founder.

Hitz

The world of data management has become much more challenging in the past two decades, but in terms of NetApp's business, that's a good thing because it gives plenty of room to innovate, says Dave Hitz, executive vice president and founder.

How has the storage/data management business changed since the time you founded NetApp?

Twenty years ago, data was technical, boring, and mostly unregulated. It was Fortran, C code, schematics, budgets, and the like. Today it is personal. It is baby photos, love letters, medical records. As a society, we haven't even come to grips with how to deal with this new kind of data. Whenever you see a headline about the European Union suing Google or Facebook, the underlying issue is usually about what we think the rules should be for people's personal stuff. Data has gotten much more interesting.

What were the challenges then and what are the challenges now?

A big challenge is that data sharing has become more common. Data is often networked, sometimes even part of a large cloud infrastructure. So sharing is easy, which is good when we want to share, but challenging for data (like medical records) that are illegal to share. Using virtual machines also makes data management trickier. What's so cool about virtual machines is that they can so easily jump from one physical machine to another, or maybe from one data centre to another with hybrid clouds, but that only works if the data can jump around just as easily as the VM. Bottom line, the world of data management has become much more challenging in the past two decades, but in terms of our business, that's a good thing because it gives us plenty of room to innovate.

How do you see the data management business evolving in the next five years?

I believe that data needs to become more infrastructural. It needs to evolve the way that the networking industry has. Once upon a time, people ran networks in silos. That sounds crazy, but if you go back far enough it's true. DEC had DECnet, IBM had SNA and Token Ring. Every server vendor had its own network. But in the 1980s, when people started installing PCs everywhere, they needed a network that would connect everything up. You can't have client/server computing if the clients can't talk to the servers.

And so networking turned into a shared resource - like a utility, managed centrally, but used by people all over. We believe that data must move in the same direction. To understand the key characteristics of a data infrastructure, think of things that are true of a network infrastructure. Once you install a network, it should live forever. It should work without disruption even as old elements fail or go obsolete and new ones are added. It should scale from tiny to giant, as your company grows and as Moore's law makes everything bigger and faster. And it should be smart enough to automate away many of the management decisions.

1 2 Next Page

Sign up for Computerworld eNewsletters.