McKeown is known for developing software-defined networking (SDN), which challenged the vertically integrated approach to the switch-and-router design of the past 20 years. At Stanford University, he leads a research group—including SDN’s inventor, Martin Casado—focused on SDN and shaping the architecture for the Internet of the future. He will be a plenary speaker at OFC/NFOEC 2013, 17-21 March in Anaheim, Calif., U.S.A.
Q. Did you face skepticism for SDN?
New ideas are always questioned—we dealt with intense skepticism for several years. This is the classic case of the innovator’s dilemma: The incumbents just couldn’t afford to see its potential.
Q. What have you learned from your entrepreneurial experiences?
I co-founded two companies, Nemo Systems and Nicira Networks (co-founded with Casado and Scott Shenker), which were purchased by Cisco Systems and VMWare, respectively. I have learned that, in order to create a successful venture, you must choose a problem that: 1) is intellectually interesting; 2) will change the practice; and 3) is not favored by industry (yet). If the industry loves it, they’ll do it themselves—they have more resources and a greater urgency than you. Next, you need to spend time actually bringing about the change in practice, which could include giving ideas away (I don’t patent anything) or creating a company to make the ideas easier to transfer.
Q. What can improve Internet architecture?
Internet architecture has been essentially unchanged for 30 years. This is in part because of the great vision of the original architects, but it’s also because the Internet is difficult to change. If we can programmatically control the network in an open and standard way, then the Internet can and will evolve rapidly. Once it’s possible to quickly introduce new functionality, we will see a thousand ideas bloom, and the good ones will prevail.
Q. What topics will you cover in your OFC/NFOEC plenary session?
I will focus on the pressing need for companies that build transport equipment or networks to invest in SDN so they aren’t left behind. At first blush, there is no change of functionality with SDN—functions have simply been moved around. In practice, however, it leads to profound changes to the way networks will be built and operated. SDN will eventually simplify and improve every network we build and use: WiFi, home networks, WANs, enterprise networks and optical transport networks.
Brielle Day is OSA’s public relations specialist.