 From San Jose, in the heart of Silicon Valley, extracting the signal from the noise, it's theCUBE, covering OCP US Summit 2016, brought to you by OCP. Now your host, Jeff Frick and Stu Miniman. Hey, welcome back everybody. Jeff Frick here with theCUBE live in San Jose, California, San Jose Convention Center for Open Compute Project Summit. Stu, I'm getting tired after a couple days of wall-to-wall coverage. Stu Miniman here with me from Wikibon and we're happy to introduce our next guest. I think first time on theCUBE, Bob Sherwood, CTO Big Switch Networks, welcome. Thanks for having me guys. All right, Rob, so what do you think of the show so far? This has been great. So I've been to all the OCP summits. This one adds even onto this. There's a lot more energy that I've seen. The big news is Google has joined. I call that critical mass between Google, Facebook and Microsoft. This is clearly becoming the place to be. And it seems like in kind of the talk in the hallways is that the Facebook kind of dominates as our event is going down a little bit as kind of the rise of the community comes up with some of those new members. And I think that's very much intentional. I think Facebook wanted to shepherd this enough to make sure it had momentum. But now that it does, I don't think they want to try to own it. And I think that's kind of the right thing on all sides. So Rob, you've got a lot of history with kind of the open networking groups. I've seen in a lot of the shows the open networking user group, obviously here, you know, open stack of the like, give us kind of your view. Where is kind of the open fitting into networking these days kind of in general and maybe specifically some of the stuff you're seeing here at OCP? Yeah, I mean, I think what's really interesting is people talk about commoditization coming from the bottom up. Open compute is the perfect example of that. What do you commoditize first? The hardware, right? They have the racks, they have the source, they have the compute. Now they have the servers. You're starting to see more and more software. So previous to this, there was an open source bootloader, there's the open network Linux project that we run the base offering system. Now newly announced there's this Microsoft Sonic that goes on top. I think you're starting to see like this commoditization layer just kind of up and up and up. And it's really interesting the effect that it's having in the industry and the attraction is getting to customers. So what do you think of Sonic? I mean, that's big push, networking, big discussion here at the OCP show. You know, why does Microsoft have a right to help, you know, drive some of that? And you know, what is big switch thing? How do you fit into the discussion? So full disclosure, we've been working with them for a while on this, so I'm pretty pro Sonic. But I think, so from a technical standpoint, I think it's a good first step. I think that the Microsoft people would admit to that that say they use it internally. It serves their purpose. It's not going to be everything to everyone and it's not going to, that's not the intent. But it could, you can kind of see down the road that that might be the case. Like, you know, this is Linux, the Linux kernel is circa 1995. This is a good first step and there's going to be some interest and then it'll be interesting to see how the community piles on and runs with it. For why Microsoft's doing this as opposed to taking off the shelf stuff, I think that just makes a huge amount of sense. The stat I've heard is by 2018, the expectation is that Microsoft, Google, Facebook and Amazon will be 25% of global network spend. It is just pure commodity of scale for them. If they can save a couple hundred dollars per switch or if they can fix a bug a little faster or ensure there's one fewer downpine, this just all makes sense for them. So Rob, what does this mean for really your network or cloud operators, you know, that just change in the network? So what's really interesting with the advent of things like Open Compute, you can actually now go through a standard build versus buy discussion for your network in a way that you never really could before. And so now these cloud operators are actually deciding, do I build pieces of my network and if so how much of this stack or do I continue to buy it off the shelf? And my claim is that for the vast, vast majority of people, buying your network is still the same, the right thing. Now there's a lot more options there that this whole hardware software desegregation has allowed upstarts, like big switch networks to swoop in and provide some interesting differentiated products. But you've also ended up with this space where if you do have a really big company and you're willing to invest the resources and you can see the return on investment, it's actually possible now to build your own networking stack. And there's a lot of open source components that are floating around and maybe some of them are right from you or part of some of them are right, but it's much more like the server world in that respect and that's kind of interesting. So I wonder what your thought is on just the conversation of transforming networking. A few years ago it was like, okay, it's going to be SDN, SDN's, the change Cisco and VMware kind of dominate a lot of discussion. You guys have been through a lot of that. Is SDN discussion? Is it the full stack and cloud? What's really important to change in networking? And what are you guys doing from a big switch standpoint? Technically speaking, all of this commodization of the hardware of the network makes it really easy to come up with new interesting architectures. And I think while I'm still very pro SDN, I think of SDN as a technology and not a product. And so it enables us to do really interesting things like the big cloud fabric product that we have that has massive scale of potential in OpenStack or enables to do things like, we'll have a talk later here about a massive scale monitoring deployment we do with our big monitoring fabric product. Those are still really interesting uses of SDN. At the same time, if SDN isn't your thing or you're not even sure what it is, what's nice about what OpenCompute is doing is they're putting all the building blocks out there for people to do what they want. And while I do believe that SDN will still be the emergent technology in the winter, now you can actually buy the hardware totally independent of that bet. So you can buy the hardware, put SDN software on it from Big Switch. And if that doesn't work for you, you can roll that off and put an open source, you can put in something from other commercial parties. And that's just a very different networking experience that people just aren't used to. So give us the update on Big Switch. How should users be thinking, when do they call you guys? How do you fit into this very diverse ecosystem that's out there? So we provide what I think of as the iPhone experience for networking, which is simple to use, you don't have to be an expert, there's lots of interfaces to it, it's extensible. If you are not in the top 50 of IT companies in the world, you shouldn't build your network. You should go to a company like Big Switch. And if you are deploying OpenStack, if you are deploying new VMware, we've got a great product for you with big cloud fabric. If you are trying to do more to monitor your network, if you're looking at, frankly, like a gigamon purchase in the near time, you should look at our big monitoring fabric product. That's really where we come in, but those are kind of little niches in a larger ecosystem. And they're very profitable niches and that's why we're working on them. But it's interesting to see how the larger OCP is enabling all of that and making it, frankly, just more acceptable for customers. All right, so if it turns into an iPhone experience, do we get the apps? I remember at the Open Networking User Group, many of the big financial guys say, I want to have a software ecosystem where I can buy network services much easier. Is that starting to emerge? We're starting to see pieces of it. I actually think much more likely than kind of mix in apps that do different packet forwarding things. You'll have mix in apps that do different things for monitoring, things that will do reactive bits. So like a fun data point I like is in the monitoring product, because we expose this API that you can have apps for, you can have something like an intrusion detection system which has kind of a broad view. Once it detects a problem, make a call into our monitoring product to say, all right, there's this one thing here that's suspicious. Let's start recording that. Let's double click on that. Let's put that into the DPI box. And so you can have this programmable, back and forth adaptive system in ways that you've never could before. And that's the things that gets me excited. All right, so yeah, security is actually something we haven't discussed a lot about at the show. What's the security angle? So our big monitoring product has this mode called inline, where it actually goes into the DMZ. And you can actually do service chaining through our firewalls and load balancers or whatever else you have in your DMZ there. And that integrates with all sorts of interesting products, you know, FireEye and what have you, allows people to roll in and roll out services better to load balances across services to, as you're trying to, so a big problem that people have is you have an old firewall, you know it's old, it has 100,000 rules on it. In some cases, actually 100,000 rules. And you want to slowly migrate to your new shiny new firewall. But you'd rather do that in a way that doesn't cause an all or nothing downtime. And so with our product, you can actually slowly start peeling rolls off and testing and quickly switching back. There's a lot of value there. Okay. Any customer stories you can share with us about, yeah, I guess, OCP specifically, if you have. So later today, we're going to be talking about a tier one financial provider that we can't name. But you'll see we actually to play our monitoring product across 120 racks in a new data center of theirs and provide what we call a tap every rack experience. So programmatically from any point in their 120 rack data center, they can pull traffic back to any number of tools that they have to do, intrusion detection, to do performance analytics, or just do basic debugging with Wireshark. And that's a huge use case for them. All right, Rob, we're getting to the end of our time. You've been coming here years and years and years. What are you working on for the next six months, 12 months, next big mountain decline? I think the next big mountain decline is to get all of the hyperscale folks working together rather than working things independently so that the rest of the community can benefit. So every time that Google and Facebook and Amazon or Microsoft buy the same switch, it's cheaper for customers to buy that same switch, probably with different software on top. So you expect to see Amazon showing up here next year? I let it slip, that's certainly my hope. All right, Rob. Rob Sherwood, CTO of Big Switch. Thanks for stopping by theCUBE. Thanks for having me, guys. Absolutely. Jeff Frick with Stu Miniman. We are live at San Jose, California, the Open Compute Project Summit 2016. Going wall to wall day two. We'll be back with our next guest after this short break. Thanks for watching.