Dan Talayco won the SDxCentral nominations for Unsung SDN Hero, and we thought it would be a wonderful idea to have a chat with Dan to get some of his thoughts around the nomination, his work on software-defined networking, network virtualization and OpenFlow, and exciting projects he’s working on. We managed to catch up with Dan last week and he was kind enough to answer our numerous questions.
SDxCentral: How do you feel about having won the nomination as the Unsung SDN hero?
Dan: Surprised, as I don’t really think it’s deserved, and happy for the attention.
SDxCentral: When did you first join the OpenFlow and SDN movement?
Dan: I started working at Stanford in April of 2009. I was hired into Nick McKeown’s group as a research software engineer to port OpenFlow to some hardware platforms.
SDxCentral: What have you worked on since then?
Dan: Indigo was the main project that came out of my work at Stanford. That’s an open source port of OpenFlow onto hardware platforms.
Shameless plug: That project is getting revamped to Indigo-2.0 and there will be announcements about that soon. In the meantime, some teasers around Indigo-2.0:
- The Indigo 2.0 architecture will be published in a couple of months
- Reference implementation might be available sometime in Q3 of this year
- Will provide a strong programming model for OpenFlow
- Support for OpenFlow protocol versions 1.0 to 1.3
- And aims to support a wide range of switch platforms
SDxCentral: That sounds interesting! We can’t wait to see more around Indigo-2.0. So, what else did you do while at Stanford?
Dan: When at Stanford, I was active in the weekly meetings where the original OpenFlow specification was hammered into successive releases, 0.9, 1.0, etc.
I also worked on OFTest which is a framework and set of tests for OpenFlow switches in Python. That was work with Tatsuya Yabe, Rob Sherwood and others. That’s been used as the starting point for the definition of a set of conformance tests from the Open Networking Foundation.
A few of us also worked on a Python implementation of an OpenFlow switch called OFPS which was intended to validate the 1.1 version of the protocol. It didn’t seem cutting edge at the time, but there’s still an active discussion about reference implementations for the OpenFlow protocol.
Most recently I’ve been vice-chair of the Hybrid Switch working group for the Open Networking Foundation.
SDxCentral: What’s your day job?
Dan: Depends on what time of day. I work as a software engineer at Big Switch Networks. I also have two kids, 4 and 2, and a busy wife, so I exercise the flexible nature of start-up work to the max. Big Switch has provided amazing support for continuing this work on Indigo and related projects even though they do not directly contribute to the company’s primary product development. The vision and direction that come from its founders, Guido Appenzeller and Kyle Forster, have provided the guidance and support allowing me to do what I do.
SDxCentral: What are you passionate about in the SDN world?
Dan: Right now, I just feel lucky to be a part of it. The people I’m working with, throughout the community, but especially at Big Switch, are without parallel. It’s a perfect storm of great ideas and ecosystem turmoil. I think we can solve some real problems with SDN and OpenFlow in particular.
On the abstract side, I came from mathematics and academics. OpenFlow is very appealing to me from that standpoint as it is one of the first protocols that may actually provide an analytical basis for understanding networks at a real scale. There’s work by Peyman Kazemian in Nick McKeown’s group on that right now.
What’s most interesting to me is how these two things, the practical and the abstract, come together in SDN.
SDxCentral:. What do you think has been the biggest impact of the SDN revolution?
Dan: I think we’re just at the beginning and we can only guess about its full potential. Prior to SDN, we’ve developed networks to the point of being barely manageable. It’s been eye opening to learn what network administrators and operators do to make these things work, and they do it in a way that allows us to take it for granted. They are the real unsung heroes right now.
But I believe we are approaching limits with what we can do with our current technologies. If something like Moore’s law is to effectively apply to networks, some of those technologies need to be updated and I believe SDN provides the best starting point for those changes.
SDxCentral: What do you think SDN holds for the future?
Dan: There are so many things. On the practical side, network deployments that use resources more efficiently and are easier to maintain and administrate. New products that allow networks to scale better and provide new services. The impact of virtual networks with their capacity to represent the multiple layers of social, commercial and operational interactions will be profound.
Combining SDN and content-centric networking architectures could have some very interesting outcomes.
Then there’s the sci-fi fan in me that expects to see unpredictable emergent effects arise as networks and storage become very large scale sites of computation. (Okay, I probably shouldn’t admit that….) I’ll be happy even if that doesn’t happen.
SDxCentral: Finally, what do you want to say to the folks who nominated you?
Dan: Thanks for the opportunity to reflect on the last couple of years. It’s been great and I think we’re just getting warmed up.
SDxCentral: Thank you, Dan, for your time! And we agree–this is just the beginning!
Check out more Open Source on SDxCentral: