In The News

10 innovators changing the game for Internet infrastructure

Monday, June 18, 2012
By Stacey Higginbotham
 
The world of information technology is always changing. But over the last six years it has started to change more rapidly with the genesis of cloud providers, the growth in the number of giant webscale companies, and the widespread use of virtualization in enterprise environments. A new era is upon us.
 
In the next five years a new way of thinking about, constructing and operating IT will emerge. Data centers are no longer the size of mini-marts but instead are mega-marts like Rob Roy’s 2.2 million square foot Switch data center in Las Vegas. Servers are no longer the unit of computing, but instead are being taken completely apart or are a mere component in the new data-center sized computer, a trend being pushed by Frank Frankovsky at Facebook and at the Open Compute Foundation.
 
The walls between data centers will also matter less and less as software defined networks help create secure, flexible bandwidth between data centers and eventually continents, which folks like Martin Casado of Nicira are working on. This new era sees infrastructure as a service and the hardware becomes a fungible element, supporting a river of data and applications that flow on top of it. The U.S. government is certainly taking advantage of this shift with its Digital Government Strategy, led by U.S. CIO Steve VanRoekel.
 
At GigaOM we’ve chronicled the development of this new era in our day-to-day writing and we’ll highlight these trends this week at our 5th Structure 2012 show on June 20 and 21 in San Francisco. Most importantly, we celebrate the people who are orchestrating this change — those who are building systems to enable and take advantage of it. Like any big shift that occurs over years, there are hundreds of influencers who will play a role, but we’ve chosen these ten who are instrumental in driving the vision. And now let’s meet them:
 
The Engineer: Martin Casado, CTO & Co-founder, Nicira
 
Martin Casado, the co-founder of startup Nicira, has a hobby of running ultramarathons. He’s willing to go slow, and take his time if it means he can go the distance. The longest race he’s run was a shocking 100 miles. It’s a skill that could prove useful, as he moves into the fourth year of running one of the hottest networking startups out there.
 
Nicra, founded in 2008, is an effort to make networks that are programmable and built on commodity hardware. It’s not exactly under the radar. The company has been profiled in Wired, and The New York Times and has even been a victim of corporate espionage.
 
At its heart, Nicira’s software, which runs on commodity gear, allows someone to abstract their network components from the underlying hardware and then program the network. Administrators can program how the network responds to certain events, where traffic flows and a variety of other scenarios — all without hiring specialized networking engineers and pulling cables.
 
It’s a little ironic then that Casado’s primary passion isn’t in networking. While Nicira could utterly disrupt the networking industry, Casado is focused on distributed systems and is looking to create an entirely different type of IT infrastructure built for a cloud and webscale world. Networking ties computing together as well as computing to storage, and until that changes we won’t be able to unleash this broader movement, says Casado.
 
Like others we have profiled in this series, Casado views the datacenter not as a collection of servers but as a computer, and a unified system. Then he takes it a step further: he views it as an entire, complicated, distributed system. “It’s not just software defined networking, it’s an even larger movement which is software defined infrastructure,” Casado says. “We need to be able to write software that provides the types of guarantees that we expect from physical gear but still allow for a flexible infrastructure on top.”
 
That view of computing is one he developed when he was in college. Casado was a CS major concerned with distributed systems and he went on to work at IBM developing the Blue Gene supercomputer where he was responsible for networking. When he went back to school he realized that the information technology world was changing and the problems of virtualization and the needs of webscale computing were familiar from his days of thinking about distributed computers.
 
At that point he was far enough away from the starting line to feel winded but still a ways from the finish line. In fact he almost took a different path. He had an offer for a tenure-track position, but after having launched one startup while he was at Stanford, he decided to make a go again at being an entrepreneur. And so Nicira was born.
 
It’s not entirely a surprise he turned academia down. His first startup, which he built while still at Stanford, did well. He developed technology they could determine when an IP address was being hidden by proxy and where the computer was actually coming from. Ten months into the startup, in the midst of a funding round, he was offered a buyout. He took it. With that experience under his belt it’s no wonder he wanted to try the whole thing again.
 
And now he’s built up Nicira to bring his vision of software defined infrastructure closer to reality. Like his ultramarathons, guiding Nicira through this transition will take time and effort, but Casado at least has a clear vision of where he wants to go.