 It's theCUBE. Here is your host, Jeff Frick. Hi, Jeff Frick here with theCUBE. We're on the ground in the heart of Silicon Valley, the Santa Clara Convention Center, at the Open Daylight Summit, the second Open Daylight Summit that they've had. We want to come down and get a feel for what's going on, share it with you, and we're excited to have Pradeep Sen, CTO Network Function Virtualization from HP, welcome. I'm glad to be here. Well, I'm glad to have you. I'm a shout out to Sar. We're getting all of his great people to come on board. Sar's a great CUBE alumni. So you're giving a keynote. What's your keynote all about? So my keynote is about the relationship between SDN and NFV and how NFV is really the use case for SDN. SDN has been around for a while, but has had a little bit of a problem getting traction in the carrier space. And now with the advent of NFV, it's become very obvious why we need SDN. So I'll talk a little bit about that and what we need from both communities. So I'll unpack it a little bit here for the folks that missed the keynote. What's the connection? Why are they so important to one another? Yeah, so with the advent of NFV, when you want to move functions around, instantiate them in different places, do all this dynamically, if the network doesn't move with moving the function, then you failed in your end-to-end service. So the network has to move with the function and how do you do that? The only way we know how to do it today is through the use of SDN techniques by dynamically changing the connectivity underneath. So it's using the programmability that SDN gives you, the programmability of the network and applying it to the functions themselves, the network functions themselves when they move around. So is this potentially the kind of the breakthrough use case that's going to help SDN kind of propel it forward? I think so, I think so. This is really what's happening. Some of the carriers who've been involved in SDN space for a while now are realizing how important this is for the NFV story and people like AT&T have already lashed onto it and are using that in the approach in their design. So yes, I think this is the use case retro launch. So in the telcos, when they have these huge systems, huge infrastructure, huge amount of traffic, introducing new technology into that infrastructure, it's got to be difficult. So how do these new things start to make their way into these massive scale production systems? Yeah, so that is a bit of a challenge because as you know, most carriers, the huge networks, they're not going to throw it away and start something new, they'll be a gradual transition. So you need some things which bridge the gap. And this is, it's great that we're doing this interview at the ODL Summit because ODL is one of the mechanisms which allows this to happen because the ODL controller is a framework. The whole ODL project is building a controller framework which allows you to control new things, open flow-based things, as well as talk to legacy networks and builds abstraction on top of that so that the service layer, you don't really care too much about what's underneath and you can bridge this gap. Obviously, you can do more things with the new network elements and the new control capabilities which you couldn't do with the old, but at some level of the service, you can stitch them together. Right, and the whole open source kind of ideology gives you both kind of a standard without being a standard, right? It's not a top-down, dictated standards policy, but it's a community-generated standards. Absolutely. But it also gives you a platform for innovation. And how is having an active open source community really help drive the pace of innovation? We, some of us who launched the NFE movement a couple of years ago, started off with looking at how to standardize things, standardize in the sense of building a common language so that people knew what we were talking about and then seeing what need to be done. But very quickly, we realized that to get to implementation and to accelerate this and not go through a typical standards process, we needed a way to implement this faster. And we latched on to open source as a way to do that. In fact, we launched an open source project called OP NFV, Open Platform for NFV, to do precisely this, which takes in things like OpenStack, things like Open Daylight, puts them together for NFV. So this whole open community approach has been very, is very critical for us, to be able to use this wide development community, diverse development community, and build on the innovation which is coming from there and put things together for purposes we need. Right, which is great, because a lot of the common story narrative is that, the virtualization getting to the network is the final piece of the puzzle. It's got to the compute, it's got to the store, and now it's really needs to get into the network to enable these application-defined world that we're rapidly moving into. Right, and in fact, we actually think of it as a few steps beyond that. That is, we've got virtualization, bring it to the network, but we think the real value starts to happen when you actually cloudify the network. That is, bring more cloud techniques in, and then start to do things like actually make the applications, the functions, not just virtualize, but actually natively cloud-ready. And that's a stage we call decomposition. So we have this virtualize, separate virtualize, cloudify, and then decompose. And in the decomposition phase, we're saying a lot of these simpler services move down into the infrastructure so that the applications and functions can become simpler and you can compose them faster, more rapidly, more innovatively. So you can develop new services, change them faster, and get them to the consumers faster. And move them, right, and move them around. That was the big theme at the docker show, right? Develop it on your laptop, move it to Amazon to do a little test dev, works great, move it into production, move it into big production. But to easily do that is a lot of infrastructure that's got to know how to move with it. Right, absolutely. And the movement, the other services you need, which tie various things together, Open Daylight is a good example of a framework which allows you to do that because it's built on modularity, on microservices, which you can stitch together. And that's where we have to get to with the whole energy story and the SDN story to be able to realize the full value. All right, so what's on the priority for the next six months? What are you working on? So these things that I just talked about is really getting the whole SDN controller bit and the orchestrator bit together, working more seamlessly, and then moving ahead to try to get into this decomposition phase. How do we build microservices, which can be used by the application? Excellent. Got your hands full. Yeah, absolutely. All right, Pridip, thanks for stopping by. Thank you. I'm Jeff Frick, we're at Open Daylight Summit. You're watching theCUBE. Thanks for watching.