 Live from San Francisco, California, it's theCUBE at VMworld 2014, brought to you by VMware, Cisco, EMC, HP, and Nutanix. Hi and we're back. This is Stu Miniman with wikibond.org, one of the biggest infrastructure cloud networking and virtualization shows. It's VMworld 2014 here in Las Vegas. Joining me for this segment is Balaji Siva Subramanian, who's the Director of Product Management with Cisco. Balaji, thank you so much for joining us. Your first time on theCUBE. That's right. Thank you very much. So great. You know, we always really love to be able to just dig into the conversations, talk about what's really going on, and of course networking's been critically important, and even more so over the last couple of years in the VMware ecosystem. So first of all, can you give us a little bit about what's your role at Cisco? Yeah. I'm going to introduce all the Cisco's virtual networking solutions, so that those include the Nexus 1000 V, but most people know about it. And then the newly introduced Cisco application virtual switch for the ACI infrastructure. Yeah. You know, how long have you been with Cisco? I think overall about 12 to 13 years. So you know, I'm sure you have, like I have, you look at networking inside virtualization. Boy, it's come a long way. Yes. You know, the V switch has been there a long time, but I listened to my friend Greg Farrow says, you know, it might have been called a V switch, but boy, it really looked like a nick from a networking standpoint. So they went to the virtual distributive switch, of course, the API that allowed for the 1000 V allowed for a lot of growth in there. And you know, both VMware has matured, Cisco's done a lot, and everyone around that has done a lot. I'm sure you'd agree. Yeah, absolutely. I think when we started this original Nexus 1000 V, you know, virtualization was just starting. And people used to manage physical ports, and now they have lost their visibility to the ports that the VMs connect into. They want to configure the QoS, the VLANs, and such things like that. And they lost visibility to that. And so they asked Cisco to say, you know, I used to manage the, you know, I manage your catalyst port at that point, and I want to manage the virtual port the same way. Give me a tool to do that. And that's how the Nexus 1000 came about. You know, today we have about 10,000 customers on Nexus 1000 V, on V-Sphere and Hyper-V. All right. So, yeah, lots I want to talk about there. But first of all, I mean, the 1000 V started out only for VMware today. It's a, you know, multi-hypervisor environment. Bring us up to speed on the kind of the breadth of the portfolio. Yeah. So we started, you know, we launched actually Nexus 1000 V six years ago in the VMworld. And now we have been shipping, obviously, for six years. And we introduced Hyper-V support in last year, last June, last decade, we announced Hyper-V support, which has been shipping for a year now. We have about 2000 customers who are either trialing it or bought it and whatnot in the last year. And then we also introduced the KVM support as well. So KVM support has been shipping since March of this year. We did a GA of that release last month for Ubuntu, and we hope to get it for Red Hat as well. So it's basically, if you look at it, we're sort of following what the customers are doing. Initially it was all VMware, you know, hypervisor virtualization. Now people are starting to migrate to other hypervisors. So we sort of try to follow the market. All right. So when I hear KVM and Ubuntu, I have to ask, is there a connection with OpenStack then as to how this fits in? Yeah, absolutely. So the KVM solution is purpose-built for OpenStack only. So basically we're not even selling it as a standalone, you know, wish-witch. That doesn't really have a lot of value. So we have worked with Ubuntu, Canonical, and basically automated every installation of it. Basically, if you think about it, there's an OBS and then this is essentially a replacement for that inbuilt OBS there. So we are automated through the JujuCharm installation that they have. We plan to do that for the Red Hat as well. So when we do the installation, you know, OpenStack has been very, one of the hardest things to install, right? And people spend days and hours and months, whatever. So we wanted to make sure that at least our solution is fully integrated into the overall solution. Yeah. So you mentioned you're going to have Red Hat in the future. I mean, that's kind of my... That's like three months. Three months. Yeah. Before the Paris Summit. So we want to make sure that... For Paris. Okay. Great. Yeah, we were at OpenStack Summit in Atlanta. Yeah. Right. One of the questions is, you know, boy, how many different distributions do I have to integrate with? So if I get Ubuntu and I get Red Hat, you've got, you know, a decent portion of kind of the Linux community there, but, you know, it ends up being a lot of work for the vendors, doesn't it? I agree. So I think, you know, right now, we probably will stop with the two. You know, I know Suze has been some people have asking for Suze, but we spent some time on Ubuntu because they were sort of earlier to the market and so there's a lot of adoption there. Red Hat, obviously, most of the enterprise are looking at Red Hat. So we will do that. So, yeah, we probably would not go all distributions. And we have another distribution today. So I'm not sure we can keep up with that. Yeah. So, you know, when I look at when the 1000V first launched, it really was to help, you know, give a tool to those networking people who lost visibility into virtual switching to allow them same tools, same things. And it was a premium product. I had to have, you know, like Enterprise Edition on VMware and there was a license on the Cisco side. Bring us up to speed is kind of what is the core value proposition of, you know, your virtual switching and, you know, what's the go-to market between, because, you know, I think Ubuntu, you're not paying for that, I'm assuming. Right, right, right. Yeah, I think if you look at the, let's talk about vSphere for a second. So we did a model like a couple of years ago. We did a freemium model, right? So we basically give away the primary Nexus 1000V as a freemium essential edition and then we pay for the sort of the advanced edition, which has a lot of security and a lot more scale and things like that. In the OpenStack model, to your point, hey, the Ubuntu is free and everything is free. What value do you really provide? I think the primary thing is even though the OBS and things are free, sort of there's no, like, really network virtualization solution. It's not like, so I think one thing we want to walk away from is that we're not just selling a switch. We're providing a solution. So we have the first one to ship the VXLAN-based overlay solution in the market. So, you know, I think today we can support 6,000 VXLAN segments in our single VSM controller. And we also allow VXLAN to move over from, so you can actually build a multiple VSM clusters, multiple clusters, each of 250 servers, and then have your VXLAN go between those clusters. So essentially, you can build a fairly large overlay solution. So that's what we're paying for. Not necessarily for areas of V-switch and here's the, that's the difference. That's what it is. Okay, so they're actually, there's a license that goes along with that. That's right, that's right. Great, you brought up kind of VXLAN, you know, the kind of overlays and underlays have been a big discussion for the last couple of years. You know, how do your virtual switching products, you know, fit into that whole discussion? So I think if you look at it, you know, with our Cisco ACI solution, our goal is to provide end-to-end solutions, right? Basically, when you create a tenant, it starts all the way from the virtual side, you know, the segmentation sort of retained over the physical side, and not only that, we provide end-to-end visibility and performance management of that. So our virtual switching is evolving. So with our ACI solution, we are integrating directly into the, into the, through the off-flex into the APIC, which is our application policy infra controller. So the overlay, you know, what we're doing essentially is tied into the physical fabric as well. So I think the, at the end of the day, overlay is convenient. If it's just purely overlay and there's a completely underlated underlay, I think there's a, there's a problem. It's not deployable, in my opinion. I mean, it's deployable obviously, but, you know, will production applications be able to do it unless you get a lot of feedback, back from the underlay? That's why I think VMware was even trying to, you know, do some partnerships with other vendors. But essentially, you want to have a full end-to-end visibility. You just don't want to just overlay and then hope that underlay works. You know, so what about how, how, how the 1000V and everything fits into the whole cloud discussion? I mean, it started really kind of local, but, you know, it's getting more and more distributed. And, you know, I'm assuming that's a big part of the discussion. Yeah. So one of the solutions we are introducing, it's sort of related to the N1KV technology, the same team is building it. It is called the Cisco inter-cloud fabric. So basically, you know, we are basically building, allowing customers to go to any cloud providers. So we have, we have announced recently, we are going to Amazon as your, or you go to BT or Portulacom or Virtua Stream. So we allow that hybrid, we allow the workload to be migrated from your private cloud to any of those clouds. So some of the technology, like the VSG, the virtual security gateway, the CSR are all going to be available there. So we actually do the virtual networking, similar to what we do in the private cloud. We, if you just move the workload, and then if you don't get the same level of support, security gateways and other things in the public cloud, you're going to be in trouble. You're going, you know, you want to be able to have the, otherwise I won't feel comfortable going there. So with the inter-cloud fabric, we are essentially taking some of our components of VSG, for example, into that public cloud, any public cloud. Okay, so did I hear right then, do you have hooks into Amazon then to be able to do that? Amazon, Azure, and the BT will be live in, I think we are in beta right now, will be, I guess, going live in about a month or so. So yeah, that's the inter-cloud fabric. So VSG is obviously investing a lot of money into building our own fabric, either working with our partners or ourselves, and we will also be able to ram to that. Think of it as this product as a ram to any services, public services. So we want to provide customer choice. Okay, so you said you want to give customers choice. Obviously, since the NYC acquisition, VMware's added a lot to their networking group. We're going to have Martina on on Wednesday. Talk about, you know, they took the OpenV switch and kind of made their own version that's inside. You know, how are you dealing with that? Does it have to be all Cisco or all VMware? Are there ways that they can play together? You know, and customers are kind of caught in the middle. No, yeah, this is, I wrote actually a blog about it a couple of weeks ago on this topic, which is that, I think because of the virtualization of becoming so important, the hypervisor, obviously it's really important that the virtual switch in the hypervisor becomes a control point. And if a specific vendor kind of locks it out, I think it makes it difficult for customers to have a choice in the solutions. Right, the good news today is that VMware is allowing Cisco, IBM, and HP to provide the V-switch solutions. You know, Hyper-V, again, is open and providing the different V-switches to be provided. Cisco N1KB is there as well as NEC, I believe. And then obviously in the open source hypervisor, either Zen or KVM, you have OBS, right? So, you know, we are contributing heavily into OBS. So, as long as this choice remains open, I think we are good to go in terms of openness, in terms of giving customer choices. Yeah, I pulled up your blog, I was looking, it looked like Juan Lodge had contributed. We were trading tweets during the keynote this morning. So, you know, it's interesting. I remember when the 1000V first came out, I talked, I was working for a storage company at the time, talked to all the networking companies, and they said, we're never going to create our own virtual switch, because Cisco's proprietary, even though it was an open API for their doing it, and I look at it now, you listen to your blog post, IBM has a virtual switch, HP has a virtual switch, of course VMware has what they're doing, NEC, you know, there's so many out there. So, you know, wouldn't it be easier if we could all just get to a single virtual switch? Or, you know. Yeah, so good question. There's two ways to do it, right? One is a native way switch provides all the APIs that every other solution needs. For example, you need open flow for some controller, you need open flow support. You need off-flex, for example, for ACI support. You know, if the native way switch provides that, I think we are okay. And at the end of the day, we don't want to be saying, you know what, let's say Hyper-V, for example, VMware Microsoft has said publicly as well that they're going to implement off-flex on their Hyper-V switch, which is awesome. So, when you have that, then we can Cisco customers or customers who likes ACI, for example, can go deploy Hyper-V workloads and get the policy support into invisibility in that sense. So, as long as the choice is provided, I think you're right. You don't necessarily have to create your own versions of it. But I think, you know, every company is different, right? They don't necessarily want to invest because some vendor wants to do a solution based on that. So, they say, okay, at least I give you a choice. You can go create it yourself if you want. I give you the APIs and you go create it yourself. I think those are okay, you know, this is fair, right? Because you can't expect everybody vendor to, everybody Hyper-V is a vendor to sit and do work for somebody else. All right, so Balaji, you know, when I think about all the administrators coming to this show, you know, three quarters of them don't want anything to do with networking, you know. How are we doing as an industry about making it easier for the administrators to get their jobs done? And, you know, last point I want to let you take is, you know, what can we look forward to, you know, that the users can see, you know, some light at the end of the tunnel to just make their jobs much easier? Yeah, I think the silo is going to grow away. I think, I really don't believe, you know, that any vendor can survive just for providing one, like storage only, networking only, and virtualization only. And that's why you see everybody's encroaching into every other space, right? So there's, you know, because the thing is at the end of the day, people don't have resources to do that kind of silo. And it takes time to put together, six months to put the solutions together. So I think that at the IT administrator perspective, you should be willing to go and learn different technologies. If you're a VMware admin, you need to learn how to know networking and then storage, and same vice versa. When I used to be a Cisco guy, obviously I used to work in Cisco Tech, so I know CLI very well. But then I started learning Ubuntu. I can install OpenSack today, you know? So those things, I have to, you know, upgrade my skills, and I'm sure that, you know, people who work attending this show has to figure out a way to kind of become this single IT person. All right, well Balaji, we're all for busting through the silos, helping customers take advantage of the newer technologies that are going to make their lives overall, make businesses more productive overall. Thanks so much for joining us on this segment, and Cube's coverage from VMworld 2014 will be right back after this quick break. Thank you.