 Live from the Austin Convention Center in Austin, Texas, it's theCUBE at Dell World 2014. Here are your hosts, Dave Vellante and Stu Miniman. Okay, we're back here at Dell World 2014 in the Austin Convention Center. Stu and I have been going all day, we'll be here all day tomorrow as well. Tom Burns is here, he's the Vice President and General Manager of Dell Networking. Tom, welcome to theCUBE. Thanks very much, it's great to be here. Networking, lots of changes coming, right? The compute scaling out, the storage is scaling out. Is the network going to scale out? The networking is absolutely scaling out. We're having to meet all the various needs of our customers around cloud, big data, analytics, mobility, security, and all this is demanding much higher bandwidth, much better latency and performance, and the capability really to give customers the choice to maybe start in the mid-range and scale up over time as they add additional applications, users, and so forth. I mean, you're seeing some real interesting shifts in the networking space. You've got a dominant player in Cisco, has 66% of the market, and then all of a sudden, VMware's making some moves, and then it's just a really interesting dynamic. I talked about the scale out. It feels like there's a little earthquake happening. What is happening? Networking is cool again. For so long, there was not a lot of change. Cisco was dominant, you had the seven dwarfs, yet you were really basing all the technology on traditional proprietary hardware and software, hundreds of protocols that no one really used. But you know that the world is changing both coming from the various things from big data cloud, like we talked about, but also the technology that's underlying is changing it. And I think what's really cool at Dell, we embrace open, standard space technology. And if you look at the network, what's really happened, this is all available. Third-party silicon, we're starting to see third-party software, common tools for orchestration and management that really help the company look at the network similar to how they look at compute and more and more at storage, agility, flexibility, and really being able to point at their users rather than managing a black box. So I think it's a great time, and Dell is absolutely embracing this transformation. Yeah. Tom, I don't know if I can say that networking's cool yet. It's definitely interesting times. It's no longer just about Cisco. It's my generation's. The line between virtualization. Networking database are cool too. Yeah, yeah, yeah, come on. You know, the physical and virtual world. And I want to dig in a little bit more to the open networking, because that's definitely interesting and a hot topic. You guys have partnerships with Cumulus. You're opening up to kind of create an operating system. Can you unpack that a little bit for us? You know, what does open networking mean? What should customers be thinking about? And what's the offerings that Dell has? Well, sure. You know, getting back to how the technology evolved. If you go back two decades ago, you had the mainframe computer. Proprietary hardware, proprietary ASICs, proprietary software, challenging orchestration, very expensive management tools, right? And so companies such as HP and Dell with the help of companies like Intel really pretty much standardized that hardware platform, incorporated third-party software, proved out the scalability and the performance and the support, and really transformed the compute industry to where Dell is obviously number two in the world, number one here in the US and in Asia-Pac. We believe that change in technology to third-party ASICs, to open available Linux-based software, to more common orchestration and management, the time is now, it's there. And so we made this announcement about a year ago. Our first contract and partnership was with Cumulus Networks. We then signed in April an agreement with Big Switch. They're kind of sitting in different paths of SDN. If you look at Cumulus kind of from a programmability and maybe virtualization, and helping companies accelerate in that particular path of SDN. Whereas Big Switch is more traditional open flow, but offering a great service with their big tap operating controller, really looking at application monitoring. Again, virtualizing some of the products coming from Gigamon and Nixie and so forth about applications. So I think the underlying basis of open networking is the tradition of Dell. Make the complex simple. Standardize on commodity-based, commercially available hardware. Offer customers a choice in software and then wrap that with great consultation, great support, great logistics. And we've seen tremendous traction, and I must say all incremental traction. These are accounts that simply are looking to make a change from networking. They're not just looking to make a change from Cisco as an example. They want to look at networking in a much different way, get everything closer to compute, much simplified, much more easier to manage, more flexibility and agility. Yeah, Tom, I think you're bringing up a great point. One of the challenges with SDN is, first of all, what does it mean? And it's not simple. There's so many pieces. My overlays, my underlays, how does it fit in my environment? The open networking attacks a specific challenge that customers have. I want to talk about another solution set that you guys are offering, which has been seeing revenue, and it's simple, which is the network function virtualization, or NFB. Yep, really not targeted to the enterprise, but the telco. Can you talk a little bit about what you guys are doing this week with NFB? Well, very simply, I see NFB as the SDN for carriers. Carriers have been managing very complex, very expensive, multiple types of networks to deliver their services. And more and more of this can move to an x86 environment and more virtualization. So two weeks ago at the World SDN Congress in Germany, we announced our NFB strategy, which is very similar to our SDN strategy. Open, standards-based, commercially available, but also creating an ecosystem of partners giving customers that choice. And we created basically the Dell infrastructure for NFB, which incorporates our 13G server, our open networking product, and our Dell storage and Dell software. But also today we actually announced the relationship with Brocade and Intel, where we're incorporating Brocade's Viata software onto our 13G server, powered by Intel, and we're giving companies the capability to virtualize those edge routers. And so again, it's about taking advantage of the existing infrastructure, trying to eliminate the number of pieces of infrastructure, putting more of it towards compute and really optimizing cost and focus, flexibility and agility. So what's happening at the skill sets level? Because practitioners are still heavily wired to sort of the old way. And they've got certifications and sort of set in their ways. So what are you seeing in the customer base and how are you sort of helping people reskill, retrain and think about the new? You know, it's absolutely a very important question. It's a great question and we hear a lot from our customers today. Traditionally you've needed these certified networking engineers in order to do these physical changes to the network, to manage it, to monitor it, to configure it. And as you move more to a simplified set of technologies and more common tool sets, these needs diminish a bit. At least at some of the core parts of the data center is that simplicity enters in a software defined data center. I think the answer is you reskill. I mean, these are very intelligent people that know a lot about the data center, know a lot about the traffic and the applications that are impacting the network and you reskill them to focus on users, on the rolling out of those applications. And really it's not about getting rid of CCIEs, although I would like to say that three years from now, five years from now, CLI is not needed to run networks. Let's go to a common set of tools, a common set of language in a Linux based environment and that should actually have those particular skill sets focused towards the users and not towards the box. How are metrics changing in the networking world? And what are people focusing on as cost per port? Well, I think it's latency, it's performance and its capacity. In addition, more and more power requirements, cooling, et cetera, and that's where Dell networking has done very well. We've actually capitalized on a lot of the intellectual property that our server team has done. A lot of our top of rack switches are now naturally are cooled. Our current top of rack 40 gig is I believe about half the power requirements to some of our competitors to provide 40 gig. So tremendous cost savings as you look as a TCO over about three years. So I think the focus is big data and analytics. I need more capacity. I need to have some of the security built into the network. I need to make it more manageable. If you look at BYOD, the trend that's been happening for some time, I need to authenticate, manage, deploy, and potentially cut off those mobile devices within the network. So I think a lot of the dynamics are around those concerns, certainly security being one of the strongest, but still a lot of focus on the network, that's for sure. So Tom, you also handle some of the converge solutions out there and if you look at most of the converge infrastructure solutions, the networking, if you talk about compute, storage, network, and management, networking kind of is usually the smallest piece. I mean Cisco obviously has a strong focus on networking, but as we look forward to some of these new applications, it's high bandwidth, really low latency, things like RDMA over convergent internet, Rocky, SMB3 for Microsoft environments are very important. Can you talk about kind of that line between the networking and the convergence and how that plays into your teams? Well, you know, the advantage that Dell has is having a foot put in each one of the domains, right, and networking, servers, and storage, and we work very closely together to an essence, put more and more of the capability towards compute or at least closer to the compute. I mean, if you think about the introduction of our Blade networking products when Force 10 was first acquired, that was about bringing 10 and 40 gig into our 1,000 E, putting the network closer to, you know, the compute helping with east-west versus north-south traffic, that helps performance, that helps latency. If you look at the convergence that we introduced with our fiber channel ethernet switch, eliminating a top of rack switch from a fiber channel perspective, giving the combination of fiber channel and ethernet in this same fixed RU top of rack product. Vertex, bringing server storage and networking together in a small office branch office environment. What's cool about that is not just the fact that all of it is in this kind of encapsulated product, it's all singly managed. So what we've done is we've eliminated, again, back to the complexity of network. We've simplified the networking capability. So the same, you know, compute administrators that using OME or CMC from Dell is able to configure and manage the switch. And now with FX2, as far as I'm concerned, we've taken a step ahead in the industry. We've combined server storage and networking and two RU, giving customers options of different modules. We've got an IO aggregator in the back that provides eight ports of connectivity per module from the inside, four ports externally. You can add two modules that comes in copper, fiber, or a combo of fiber and fiber channel, excuse me, and ethernet as well, and really lowers the, in essence, the wiring, almost eight to one. So if you're connecting eight, 10 ports from the server into the internal, we basically can take that out to the top of rack with one cable. So tremendous, you know, consolidation, eliminating steps, but lastly again, all singly managed by one management system across all three domains. So I think that's what's key for us. We've studied this notion of, we call it a single managed entity and how it saves time. We had Sam Greenblatt on just now and we were talking about, I was lamenting about how labor-intensive our business is, and he said, you know, I got a new BMW and you open up the hood, you can't touch anything. It used to take a half day, now they can do in minutes and they've driven the cost down. But so it's saving time, you use the automobile industry, so that's clearly coming to the IT business. What are people doing with all this freed up time as a result of these sort of converged infrastructures and the like? Well, if you imagine a cloud environment or an infrastructure where in essence, you have, you know, server storage and networking here and you have a virtualization lever and then there's all the services applications or workloads and various, you know, things that you want to deliver going back to the NFV environment, the carrier environment, obviously, delivering services to their end users and their end customers. Customers are really looking for that lower portion below the virtualizer to be, you know, in essence, maintenance and management free. They just want it to run. So they're going to divert resources to the upper layer, the upper APIs, doing those service applications and delivering more and more. It's what we call the Future Ready IT, right? The traditional IT managed boxes. The Future Ready IT has to work with its users, deliver services, has to go from doc to rack to deliver if these applications must faster and has to eliminate those steps that you talked about out of the process. So I think, you know, these people, they're focused on their users. They're focused on using the technology to create differentiation rather than managing, you know, three silos of the infrastructure and really creating competitive advantage. I mean, that's truly why I'm in this business. I'm passionate about using technology to help companies create, you know, competitive advantage. And so presumably that's touching the application more, supporting the application development, new business models and the like. Somebody remarked a couple of years ago that the hyperscale guys, the Amazons, the Googles, they'll spend, you know, engineering resource to save money, but IT guys, they'll spend money to save time. And they're looking to buy a brick of, a block of infrastructure that they don't have to mess with. And that's really what you're providing here. No, absolutely. I mean, if you think about the first concepts around software defined networking, the belief was, well, the switch is going to become cheaper. So your CapEx is going to get reduced, right? That's what everyone kind of thought. I mean, you talk to customers now, you're looking at deployments. In fact, we just had Medea, one of our customers on Cumulus and with our 40 gig. The things that he talked about was the time to reduce and getting those stages rolled out on his applications and facing his users. He's a software as a service provider. He's providing customer relationship management on demand for up to, you know, multiple customers to as many as 100,000 users instantaneously. He's got to be focused towards them. He hasn't to have to be thinking about how do I roll out the switch? How do I bring it together with my storage and my compute? He wants that to be seamless, consistent management. So, you know, SDN, SDDC, whatever you want to frame it, it's about handing more flexibility, more agility to the customer so that they can focus on what's important to their company rather than, again, managing boxes. Tom, how about quality of service as we move to this sort of elastic, pay-by-the-drink environment? What should customers expect from quality of service in terms of their ability to essentially policy it, control it, and deliver to their constituents? Well, I think, yeah, it's a very important question and I think obviously quality of service is still absolutely critical. And I think when you and I are using the network when something happens, we're upset. And so, the quality needs to be there. The question is, is how do you manage, in essence, virtualize that quality of service between the various applications, whether they be voice, data, or video, and basically change that on a real-time basis based upon demand. And so, actually, that's where we see some of the current SDN deployments. It's really in this multi-tenant environment where on a real-time basis, you can, in essence, adjust the network from a software perspective, applying more QoS to a particular department or application, maybe de-throttling certain other areas where you don't need that sense of quality of service. So, absolutely still critical from the network performance standpoint, but I think the whole area of software-defined and putting more of that control at a software level, vis-a-vis programming, a controller, or a hypervisor is absolutely critical. What about, and we got to end it here soon, but let's end on this. What about sort of thinking ahead to the future? What are customers wanting from you? What should we think about the network and the converged infrastructure going forward? What's it going to look like? What are the requirements that the industry's not meeting today that you guys think you can solve in the midterm? Well, I think we're solving some of it, and in fact, I think it's back to the overall Dell enterprise strategy, which ties in well to the Dell strategy. These are the industry trends around big data, cloud, all the things that we've talked about. How do we take an IT from the old IT to the new IT? How do we have them compete with the likes of a Microsoft, an Amazon, a Google, which provides real-time services on demand to their users? How do we create that environment? And part of that is bringing server storage and networking together, improving the management, the agility, and the flexibility, meeting the demands of their particular users, and I think that's what we're focused on. I think the other thing we'll see now, if you thought about this 20 years ago, converged infrastructure existed. It was these huge IBM machines that had server storage together, and they took up probably at half of the size of a room. If you look at our FX2, which incorporates high-capacity compute, storage, and networking into a two-RU fixed module, starting from the small SMB, being able to scale up, being able to stack, we're looking at actually making the data center smaller, easier to cool, less power, and just as much capacity and scalability as there can be, again, making it kind of that platform rather than three silos of server storage and networking. Excellent, all right, so giving the folks in IT an ability to behave like mega-server suppliers is really what the mission is. Absolutely, I think that's the plan. All right, Tom, thanks very much for coming to theCUBE. Thank you very much, appreciate it. All right, keep it right there. Everybody, Stu and I will be back with our next guest right after this. We're live from Dell World 2014 in Austin. We'll be right back.