 Live from Las Vegas, Nevada. It's theCUBE, covering EMC World 2015. Brought to you by EMC, Brocade, and VCE. Hi, everybody, we're back. This is EMC World 2015. I'm Dave Vellante. John Furrier is here. Kumar, this one up then is here. He's the chief technologist at Verizon Labs. He's up on the keynote today, Emerging Technologies Day as one of the customers. Kumar, welcome. Thank you, Dave. Thank you, John. So what is Verizon Labs? Tell us about that. Sure. Verizon Labs is part of the overall product organization within Verizon. Our goal is to look at technologies that we can build as well as technologies we can buy from our vendors that can create a common platform across all our products for us to launch products a lot quicker than what we've been used to. We've traditionally been a fairly vendor solution dependent model and that has its challenges. So we are actually going away from that model and trying to own a more common architecture that will let us launch new services and products fairly quickly. So that's what the organization was set up for and the key talent in there is software development talent that we brought in as well as software architecture and the DevOps skillset that we can bring into launch and run applications and that more. So that is really what Verizon Labs was set up to do under the product organization within Verizon. So talk about your business a little bit. The big four, cloud, mobile, social, big data. You're certainly taking three of those four boxes at least that we can talk about social, I guess, but you're at the heart of the mobile activity, very much data driven, service provider, you are cloud. But things are changing in that business. How are those changes affecting your business? What are the challenges and what are the opportunities for you? Sure, so you're absolutely right. We're in all of those businesses. We're driving a lot of the change in that business but you're also aware of a lot of the macroeconomic pressures within those businesses. There's a lot of change that is happening in both in regulation as well as competition that always continues to threaten that base. Our goal is always to look ahead to see what we can do from a technology point of view for us to be more cost effective and where we can leapfrog certain technologies to get to the next generation of technologies that will help us offer services quicker, faster, much cheaper. And that's been always a challenge because when you see the technology landscape, even if you take EMC world, the number of different options you always see is so many. So my group constantly goes through trying to figure out what are the areas we need to play and what are the areas we need to own? How do we go about creating new business from some of these? How do we optimize existing work? So yes, we participate in it but we're always constantly on the lookout to say, what should we do to be different next year? What do we need to do to lay the foundation for a completely new rise? It's a big challenge. You've got an industry that has huge capital investments. You've got to get return on those assets so you have to keep those lights going. Like you said, cut costs, everybody's trying to do more with less. So what's your innovation focus? Where is that innovation coming from? So let's take specifically the data center. So if you look at my product, all of the products and how they currently run in the data center, you would, I mean, it is typically what you would expect from a data center that was built, you know, more than a decade ago. You would have silos. Every application or service has its own silo. And I mean silo, I mean the application, the operating system all the way down to the hardware. You are completely siloed to a point where if you look at, I mean, our way to achieve fault tolerance was redundancy. What I mean by that is you have completely replicate copies of certain things running for you to be able to achieve the scale you want to achieve. But that is a downside to it. The downside to it is that I have stranded hardware. At the lowest layer, if you look at it, I have stranded hardware. My power usage efficiency is incredibly inefficient because I have two CPUs doing the job of one or maybe a fraction of it. And my operating costs and trying to keep up that many boxes, not just the people, it is the power, the cooling, everything is high. And now the headache of maintaining so many different operating systems. So I mean, I have so many flavors of Linux that are running in my operating system. Each vendor brings his or her own flavor of Linux. Now the problem is if I have a bug, like a heart bleed bug or pick any other bug, the number of different patches I need to install and how I can roll out those patches, it's not instantaneous. It's over a few weeks or months. I can never get anything down quick. And the key I want to leave with you is that if I transition that to a more homogeneous environment. So homogeneity would give me the operational efficiency I need by saying that actually you don't have that many options on hardware. I mean, it's commodity hardware. You have three or four flavors of boxes you can buy, typical server type boxes, some with storage, different types of storage, maybe some with flash storage, some with disk storage, but everything else I'm migrating to the software. So it's very easy to do on some of the technologies are more mature than the others. I can talk about, if you look at on the compute side, we want all our software, not just the software Verizon rights, but even the software our vendors bring in to completely transition to a model where it's micro software architecture components that you can bring in. They run as loosely loose components that are coupled together with well-defined APIs, if you will. So I can now take these containers as we, early on we invested heavily in Linux containers and said, you know, that is the way of the future because it provides a clean name for us to package all the way to production, how we take the software. So that model is solved. Networking, we have talked enough about software-defined networking. There's plenty of solutions for us to run very scalable networking within the data center. So applications can request access to outside networks disparate external networks fairly easily and how they get granted, what they do with it, very easy to achieve. The big challenge was storage. And the reason I'm here, the reason I was talking to EMC is how do we solve the storage challenge? Quite honestly, I think when you listen to CJ talk about in the keynote, some of the technology, the emerging technology from EMC was hard to find. We couldn't actually find SCALIO because when people thought about EMC, they were thinking about big boxes that took up the entire rack and had lots of lights and lots of colors. And that's really the vision of EMC. But when we started engaging with the emerging technology division under CJ, actually CJ made the personal pitch and the Kudos to the Verizon accounting for pushing hard to get CJ to come talk to us. But once we got that engagement going and we got a vision into SCALIO and ECS, there was almost no going back because it kind of opened a whole world of possibilities that we had not considered before because we did not have to come out of storage. So you can take some typical scenarios. For example, what happens when disks fail? Your applications still need to get running. What happens when you have networks going up and down? Still need to get running. What happens if your entire data center is isolated? I still need to get my apps running. If you look at some of the guts behind how ECS is built, a lot of the geocaching, georadiance that they build in adds tremendous value to how I can now leapfrog onto my next generation of the cloud where I have all this compute, the networking and the storage completely in software. So just imagine a container with some software running on it that requires storage. Container is running, it's running happily, it's processing its load. Whatever happens, you have to now migrate the container from one machine to another machine either on the same rack, completely different rack, a completely new data center. I can have the storage following. So the application wouldn't even know that it lost the storage because ECS would let me follow it. Same is true for Scalaio. I can just pick it up and follow it so that I have the renunciate need. So back to your original question, if I now get to that homogeneous environment, I drive down my operating costs, I drive down my complexity within the network. And the most important thing for me is I can launch applications a lot quicker because nobody needs to worry about the whole silo. My point is just bring in your components, plop it in, get up and running. And that's your scale model. That's my scale model. So you have the DevOps philosophy in the labs, silos are kind of like really not compatible with this new platform three they call it. We call it, you know, big data cloud. How are you guys enabling those apps? I mean, APIs, new CMSs are coming in, get node, Angular, new technologies. Real time is big. How are you guys dealing with that new real time piece in the static silos, firewall rules and whatnot? How do you guys deal with that? Is it software? So in the static silos, actually we have, I mean, when I talk about DevOps, I'm really talking about DevOps only in the new environment. Our existing environment, the siloed environment is actually follows a traditional vendor operating model. But on the DevOps side, if you look at what really is the key, I mean, it's true for everybody who does DevOps is automation. And the key to automation is API. So whatever APIs we can leverage. So for example, it's not that we know that a disk has failed. Before it failed, we would know how many read write errors it has to predict the disk can fail. So for us to transfer the workload. So everything about the DevOps model on the new cloud is all about what I can automate. And to be quite frank, there is no one magic bullet that solves the whole thing. A lot of that are stuff that we are building, we are growing, we are learning as we go and we are writing it. But about 70 to 80% of any DevOps is very application centric. So you need to know what your application is doing to do it. There's no turnkey boilerplate product. I mean, there's some tooling that you could use, but. Absolutely. There's plenty of tooling we can leverage, but ultimately at the bottom line, it's all about how we can get a lot more automation. So the key about automation is once I have more automation, I don't have any procedures or mobs. I eliminate all the people who have to now take it from development to operations. So when I eliminate all that, I create the seamless cluster operating system, if you will, for me to launch new applications. So Kumar, talk about the personnel changes. Obviously with automation, orchestration, this is all goodness in the cloud, and it's evolving, it's early days. But roles are changing, right? I mean, the old storage guy, provisioning the silo and statics on the infrastructure, with DevOps, there's personnel changes. How are you guys hiring, share some insight into your thoughts, your vision, and then some examples of the kind of role changes happening within the organization and to position yourself for those apps. Absolutely, and you hit the nail on the head. I think the skill set we are looking for is grammatically different. We are not looking for an operations skill set anymore. DevOps is all about being able to write tools, write scripts, write technologies that can help you drive the automation. So it is certainly new skill set. And in Verizon, we're consciously working to attract the talent that we need, and it's not easy because when you go to a college campus and try to recruit as Verizon, the first impression is it's a cell phone company, there's a lot of technology behind it that we are trying to drive. And that has been a challenge that we have been working towards. We're trying to create innovative ways, but the personal change will happen. It has to happen for us to be able to take advantage of new technologies. Because frankly, the days where you could create those silos are behind us. So there's no way we can go back. But for us to take advantage of it, we do need to get new technology. We are talking off camera. I asked you about network function virtualization. You were too sanguine about it. What's your take on NFV? So when I was talking, what I was trying to explain about that is, there is multiple ways. So the network function virtualization really came about where you took the same existing functions and just ran the network functions in a virtualized environment in the cloud. Now, what is the problem that it really solved? All it did was it let you optimize on the hardware specific to the hardware purchasing, if you will. So instead of me buying three flavors of x86 servers, one purple, one green, one blue, I'm just buying one flavor of a server and I'm running VMs and running all of that inside it. But I submit to you that the fundamental problem that we face is a software management problem and not a hardware management problem. So if you now peel back NFV and look at the VMs, the VMs are probably running operating systems that were deprecated before I was bombed. Now, how do you support it? Because it is easy for a vendor to just pick the whole software, port it, and run the entire virtual machine and say I have a virtual machine up and running, I'm NFV compliant, but it doesn't solve my problem because my software maintenance on that still is going to hurt me. So when I told you I was like this on NFV, my point is that when we need to transform a network, I need to look at the holistic picture. I need to look at how I'm solving the software problem, how I'm solving that. The hardware problem I submit is a more trivial problem to solve and that's why when we looked at the cloud, we did not just use standard virtual machines and say bring your own OS, run whatever you want. I want homogeneity and then I can drive automation and software management on top. That was why. And that's your scale strategy. All right, Kumar, I'm sorry, John, we got to go. We're out of time, got to leave it there. So thanks very much for coming to theCUBE. John, I will be right back after this brief word. We're live, this is theCUBE. We're ADMC World 2015, we'll be right back.