 Live from the Moscone Convention Center in San Francisco, California, it's the Cube at Oracle Open World 2014. Brought to you by headline sponsor Cisco Systems with support from NetApp. And now here are your hosts, John Furrier and Jeff Frick. Okay, hello everyone. Welcome back. We're live here in San Francisco, California, Moscone Center, where Oracle has taken over downtown San Francisco. Of course, the Cube has taken over the Cisco booth and we are here live. Thanks to Cisco Intel NetApp, the Cube's here. We have two Cubes, our 50-year covering it. We're also over on the Cube Logic with Dave Vellante and Stu Miniman. I love to have two Cubes. It means double the action, double the fun. I'm John Furrier, the co-founder of SiliconANGLE Medium. Here with Jeff Frick, my co-host. Howie Shue, Senior Director of Cisco Cloud Networking and Services Group. It's back on the Cube. Great to see you again. Thank you for inviting me. So I love the fact that we're in the Cisco booth because I'm a big fan of Cisco, big franchise. Obviously, they made the Internet, TCPIP, networking, routers, edge routers. We'll come back to that term in a second. We don't have to disrupt their technology with the edge means, but networking, obviously, is in Cisco's wheelhouse. Great market, great activity, a lot of change happening. A lot of people going, finally, some good stuff happening in the network, intelligence, and obviously the apps are driving it. So I want to get your take on a couple things. One, cloud is now legit because Larry says it's legit. So, you know, that's one comment. And the joke we were seeing on the Cube is if you just called network computing cloud, it would have been successful back in the day. So joking about his early venture with terminals or basically dumb clients, if you will, or mobile devices. So get your take on the current cloud situation. Obviously, we talked about intercloud earlier yesterday. Here at Cisco, you're a new foray. What's going on with cloud and why is the networking piece still the most important discussion item in cloud? Well, you know, this talk is actually at the right timing. Cisco actually announced a number of new things with intercloud yesterday by Rob Lloyd and other senior executives. In particular, there are a number of products. I actually am responsible for intercloud fabric. The idea here is if you ask any CIOs today that, okay, do you want to go cloud? The answer is absolutely. I want to go as fast as possible for many of them. But then you turn around and you talk to a CSO and the CSO is going to tell you, no, slow down. I lose the visibility. I lose the control. I lose the compliance and the security stuff. So that's the fundamental issue that cloud era is at today. Now, intercloud fabric and the Cisco's take is you don't have to compromise. We give you a fabric. We give you that intercloud technology so that you can actually go to CIO who wanted to go cloud and go hybrid cloud. We give you the technology so that you still retain the compliance, security, and the control and the visibility. And without the vendor locking because you can move your workload to one public cloud and you can move back when you want and you can move to another public cloud. So the idea Cisco is pushing for is really the intercloud notion that one cloud has its limitation. It's the ecosystem of the cloud. It's very much like an ISP for the internet. Yeah, I mean, the play on words. Internetworking, which made, was the mega trend that really made Cisco and 3Conf, that matter, when they were still around, made the networking business. I mean, back in the day, internetworking is all about sharing across different networks. Here, it's interesting. It's clouds, which is combination of networks and other stuff. So talk about the other stuff because it's not just networks anymore. There's other parts of the intercloud. You got virtualization, you got application specific workload requirements, a lot of policy. How complicated is it? And what did you guys do in intercloud to make it easy for customers? That's a good question. The reason there is a reason we call it intercloud fabric and not intercloud the networking fabric because it's way beyond the networking. Sure, you need to give people the same networking primitives or the security or the compliance. But then when you move workloads around, it's also about the storage. You have to move the storage along with you. If you don't, then your performance sucks. And then you also have the interdependency of different applications, figuring out what is the sort of the cluster of the application. Should you move together? If you move one or two components, is that going to, you know, compromising the performance? So it's actually a very complicated stuff. So I call it somewhat rocky science stuff. Yeah, so also let's ties into some of the things that we're seeing, I see with Oracle the past few years, engineered systems. We've seen things like VBlock. We've seen things like FlexPod come out. These are engineered systems, if you will. They're essentially high end conversion infrastructure platforms that have been tailor made. So it seems to be that the trend is towards engineered solutions for I don't want to use the word engineers and quotes, but like, you're seeing like, let's just say Goldman Sachs is a big company might have a requirement for certain utility in the network and the applications for transactions. Okay, real time or something of that nature. That might be a great blueprint for something else. So, so what do you guys do to enable a customer? So they're not fearful, because the fear is, well, I'm using the inter clouds and stuff moves around. How do I know it's baked? How do I know that the environment is baked out in a way that's going to support it, hit my SLAs, if it's real time, I need performance, I need late low latency. All these things are kind of the fun and the cognitive dissonance that they might have after they go there. Yeah, I think you touched up on a very important part. It's not just the technology. It's the solution that matters. When I talked about the CIOs wanted to go to cloud because they heard all day long that cloud has this benefit has that benefit. But that's a lot of that is just technology. What it comes down to is, you know, what is the security and the compliance and how easy to get it going. CIOs want something that it just works. It just a plug in the play. If you look at some of the open source project, some of them are very successful, some of them are less successful. They're less successful because people pay attention to the technology piece less to the 20 solution piece. If you look at a Cisco as a company, we are delivering not just the technology, but all of the solution. So you bring this up. This is kind of the speeds and fees. We always say that this port has X number of ports and they're certainly fast. That's the old way. The new way is business outcomes. It sounds like a cliche term, but the end of the day, the investments going into specific managed business outcomes. So with that in mind, how does intercloud play into it? Because now, do you certify clouds? Is NFV involved? What are some of the things in there cloud could fabric? Can you share some some some data? And one, do you certify? And does NFV have anything to do with this? Well, all of the above, but let's sort of decompose this a little bit into cloud. First and foremost, it's actually a collection of the clouds. There are participants like BT, Telstra from you know, outside US and the inside US. There are also partners working together. It's kind of important because a lot of the countries or the shops, they are very much geosensitive where the data is, where the application is. So that's why we are we are cooperating with lots of service providers around the world. So that's one part. Well, and that that criteria by localization is two full performance for the user experience and also compliance, right? I would say compliance terms everything else at this point of the game. Now and then you said this NFV stuff by NFV as a means to an end to your point of business outcome is what's more important. So what is the business to service providers providing networking as a service utility networking services? That's what they want for that's for the service providers for the enterprise guys. I just want to make sure when I move the workload from here to be I enjoy the same security mechanism. I enjoy the same compliance and then you have to provision. You have to deploy the networking on demand. Oftentimes you need NFV to help. So NFV is there in the picture, but it's a means to an end. Okay, so I got to ask you about SDN. So SDN since Nassirah went to VM where it seems like a decade ago. My God, it's a long time. But two years ago now full two years. I mean, dog years, what do you want to call it? I mean, but a lot's changed. How has SDN changed? Especially this past year at VM world. Martin was on the cube and he said there was a watershed moment this year in SDN that finally the damn has broke. That was his quote about customer adoptions kind of in this phase of is it real? What's happening? What's your take on that? I think for VMware specific, I cannot comment on the VMware's business because I don't have a lot of insights. But I want to say that what VMware's message NSX message resonates well with customer today is actually a technology that has been there for more than 10 years. Mendo actually did a talk on this East West traffic security more than 10 years ago. But security is actually getting more and more sort of attention because of all the things going on. And so when you talk about this networking on demand going this and that for enterprise guys as a technology solution, it's not quite ready. However, one thing that it's sort of addressing their fear or uncertainty or the further part is the security part. I'm going to tell you that, okay, you have lots of virtualized the server and the East West traffic is not quite there. That's sort of, you know, immediately raised the tension so that that's why NSX is getting some traction in this East West security business. And so NXS aside from Martins, I mean, actually he has his, you know, thing to do with there because he got bought out. But in general SDN, how would you give it a grade? Doing well, A, B, C, you know, B minus, A plus. That's interesting. You actually asked the discussion to Steve Harrow, the dance CTO of VMware about two or three years ago. I believe Steve gave computing virtualization as A grade and then storage close to A and then networking virtualization C plus. I want to say we are still at a B ish grade. The reason is if you look at the virtualization computing virtualization, it's mature storage virtualization. You look at Oracle world, you look at a VMware that's all about storage virtualization. If you look at a 20 solution for networking virtualization, it's still limited. Of course, Cisco has its own solution, which is the ACI application based on networking. And the idea is having admins or the networking guys managing one network overlay and underlay together. VMware is more pushing towards the direction that I have an overlay and then managing overlay only. But then you since you cannot fire any networking guys managing underlay, you still have to have a different set of people managing the underlay. So in that case, you have two networks. That's additional complexity. Cisco's messages, I give you one solution. And you guys will manage the overlay and the underlay and one product. Yes. Okay, cool. Well, that makes a lot of sense on coordination. How does the UCS fit into that? Because one of the things we heard from your customers is the service profile has been an interesting thing. One of the customers actually called it the soul of the network comes out of the box and can be moved across the network. How does that relate to all of this? So if I just go back to the history, why virtualization was interesting to the enterprise? Because I'm managing all the virtual entities. It's flexible, it's sort of easy to manage, provision speed, all of that is cool. So what happens is that people look at a virtualization world and then look, okay, physical world. Why physical world is so inflexible, so clumsy? So the first thing Cisco did it was, okay, let me look at a service. Why server provisioning is so so difficult? Because you still manage server like a brick, it's not intelligent. Now I give you the profile, I give you all the intelligent tools. So now you manage servers, the bare metal servers. Management becomes a lot easier. So that's one thing Cisco did. So what ACI does is essentially apply the similar model to the physical network. We used to be very clumsy, slow. Now I give you the intelligence, I give you application awareness, I give you this profile-based management, policy-based management. So the idea is the bare metal, servers, bare metals, networking and then in the future bare metal storage is going to be very flexible. So what's it, if it's a brick before, what is it now? Well UCS, UCS grew from nowhere five years ago to, this is, you know, number one in US play market for for good reason, which is the way you manage UCS is a very different, a very different level from the way you managed in the past. Yeah, in the past I mean the brick analogy really good means it's a dumb, dumb, dumb box. But back then you had the hard wire, you know, local networks, configuration management, all kinds of management. It was very difficult and that was just physically, now you have virtualization. Okay, so let's take it to the next level, disruptive technology. So I've got this network that's smart, got some agile capabilities, adaptive, whatever buzzword we want to use, but essentially flexible. So now you get the cloud comes in, now you got application developers DevOps pushing down saying I want to dictate to the infrastructure how to behave based on the workload or the application. But I'm a developer, I'm not a networking guy. So that's DevOps, that's infrastructure as code. What is the thing that you see happening most important in this infrastructure as code movement? Where the app developers want to provision, they want dynamic, they want policy based, fill in the blank, for all aspects. Well, so the analogy I give you, virtualization actually looked at how clumsy, how inefficient the physical world was and then took advantage of it and then created an industry, virtualization industry, right? Now people look at the virtualization industry, just like we looked at the industry about 15 years ago. Actually, virtualization brings a lot of fat, a lot of inefficient signal. Now how do I go 10x more efficient, 10x more provision speed or agile? So the container technology is obviously in a lot of people are talking about. When you look deep, I looked at the container technology, not just the Docker, Docker is great as a great company, they just raised a huge amount of money. But I looked at it as fundamentally what is the management unit? 15 years ago the management unit is the physical machine that was inefficient. And then we bring to the virtual machine and then that becomes the management unit. Now, next 5, 10 years, the management unit seemed to me that should be the application directly. I went to VMworld, the CTO did a keynote speech, in his keynote speech he said, well, do you really want to manage thousands of Linux threads? That's unmanageable. I agree and disagree with him. I agree with him that today with today's technology, it's not manageable. But I think this is where the container is going to lead us to. If container world is going to give us the framework, just like with virtualization, VMworld data to the physical world, it gave us the framework so that we can manage threads, manage application as skill in the manageable way, give you the lifecycle management. At that point, you cannot argue this is actually, you know, auto magnitude more efficient to technology at that point. I mean, the computer science innovation is interesting. If you take the container, it's got traction as an interoperable unit for the apps. You can write to it, the developer community has traction with the way they wrote their open source piece of it, the way the contribution works. You can almost bolt on a compiler to the container. You can bolt on other intelligence to manage things as complex as threads. That's what you're saying, right? Yes, and at that point, once you have that technology, at that point, you look at a hypervisor. Hypervisor is actually in the way. The reason is, today, what do you do whenever you're provisioning a virtual machine? You say, I give you 8G memory, I give you 20G disk, but look at enterprise applications. What's the last time you really have an application? You really need 8G memory, 20G disk space, 24 hours a day, seven days a week, you know, 365 days. Eight cores, add some cores on there. There's no way you need that many resources all the time on the lasting basis. When you go to the application level, container level, you are actually scheduling things at a much finer granular level, and then that's where the efficiency comes from. What's interesting, you mentioned threading. I would maybe tap into the multi-core and memory. If I have a specific app, that was threading. Multi-threads all over the place at full peak. But what's interesting about virtualization in the cloud is you can move resources around, right? So this is an interesting dynamic. The other thing is, one of the key problems virtualization industry addressed was multi-core when you had an inter-release multi-core and then Windows and Linux at that time couldn't quite take advantage of the 16-core, how many cores of virtualization actually addressed the issue. But now the problem space actually moved because application actually sometimes needs thousands of cores at one time at a scale-out way. Virtualization doesn't quite bring that benefit, and that's where the thread level scheduling actually makes sense. Okay, well we love the container technology, Yasu. I want to get your question. I want to ask you the same question I asked Steve Herrod at VMworld since you brought up Steve Herrod, which is a great friend of ours. We love Steve, always great guest, like yourself on theCUBE, very candid. And a great mentor of myself. Has he been? Oh, yes. He's fantastic. So he's done good. I love Steve. Good karma. He lives in the new karma world. Hi Steve, if you're watching, we want you back on theCUBE. Hi Steve. You owe me coffee at the new place with the whiteboard, the general catalyst. Okay, so the question I want you to talk, I want you to comment on three bullets. We're living in a multi-cloud world with infrastructure for mobile and perimeter-less IT. This is a concept we've talked to Steve Herrod at VMworld. Comment about that. What does multi-cloud world mean? What does infrastructure for mobile mean? And what does perimeter-less IT mean? Perimeter-less IT. So first and foremost, remember 15 years ago for events like this, what do we talk? We talk about CPU speed. We talk about that. And today we talk about cloud and mobile. And I think it's just the same Moore's law. It's just a different manifestation. Moore's law was sort of, you know, people look at Moore's law by the CPU speed and the things like that. Today people look at it by cloud, by the mobile. So I think it's the same stuff. Now back to your question, multi-cloud. Today, like I said earlier, people have a genuine problem. The genuine problem with the cloud is that there is a vendor locking. If I move to some of the public cloud, how can you move application back? How can you give the enterprise level of the security and the compliance primitives? So that's where the multi-cloud come into play. You have the flexibility, you have the hybrid in this, you have the vendor choice. So that's one thing. And also that's also where inter-cloud might fit in for you guys. Yes, inter-cloud fits perfectly well and then that's exactly why Cisco's motivation to address this. Just like Cisco solved the problem of internet working, now Cisco wanted to solve this inter-cloud problem. So that's number one. Mobile. Mobile is just the way people consume technologies in a very different way. Before it's the PC or one notch, now it's the iPhone. It's just a different way to consume the technology. What's important here is a lot of the enterprise guys, they spend a lot of time in the last 15 years configuring CLI. And then they look at their iPhone and then they're looking at the user interface. Inevitably, they would go back to vendors like Cisco and say, how come I don't have an iPhone-like interface when I configure switching the routers? How can this is a very much policy-based? How can I just configure switching the router as easily as... And the horsepower is not that much on the iPhone, even though it's getting more and more. So you might have a master slave architecture. The compute might be different. The low latency, real-time data. Yeah. So a lot of people talk about the mobile things. The key point I wanted to raise is that kind of user experience is actually feeding back to the enterprise. Now, enterprise guys are asking companies like HP, Cisco, or Intel, hey, you know, whatever the technology, you need to give me iPhone-like user experience. That's actually a very profound change. I'm seeing that first kind of for my own product, for my own team, you know, I need to get my team to sort of develop a product, not just for CCI-EPHDs, but also for iPhone generation. Yeah. So the last thing is the sort of the security or whatnot without permit. That's actually very much because of the virtualization, because of the cloud. Before you have a choke point, you can say, I put on my switch, I can put my firewall right here. And then I configure the policies. But these days, because of the devices, mobile devices, because of the virtual machine move around, because of the containers moving between different clouds, I mean, it's very hard for you to find that choke point and let alone having a choke point, that technology, that really scale at the first place. Okay, great. So apps driving the future. Multi-cloud, okay. IT perimeter, less security because that's the reality. Where does that leave the customer right now in your mind? So you're developing and what the customer's doing. Tie that together. For the customer watching out there, Cisco customer or IT buyer or just an enterprise who's got a business objective. What do they do? What's your advice to them right now? How do they handle all this? What do they do? Is there a roadmap? Is there a playbook? So for the IT guys, one of the things they are paying attention to is they are looking for the turnkey solution, not just technologies. If you look at why cloud is thriving today, cloud concept or utility computing concept was there 10, 15 years ago. And then you probably have talked about utility computing even back in the last two recessions ago. Why cloud is thriving today? That's precisely because the cloud, the data center technology gets more and more complex. The networking, storage, sure, virtualization makes things a lot more flexible, brings you a lot of more efficiencies. But at the same time, virtualization also brings the complexity into the picture. That's what I also view that, you know, well, virtualization solve the problem is the Microsoft scene, which is the efficiency issue. But the virtualization also brings, you know, one pinpoint, the key pinpoint to the data center, which is because of the virtualization, very hard to keep track of things, very harder to measure things. So the complexity is just, you know, incredible. Because of that, people said, well, you know, I'm not going to do anything. I just go to public cloud because these guys are going to manage resource for me. I'm just paying as you go. And at that point, cloud gets so much momentum. Of course, at the same time, like I said, CSO is saying, no, no, no, slow down. Yes, you can go to public cloud, but the security and the compliance problem is not completely resolved. So that's why, you know, cloud is getting momentum. However, the problem is still yet to be solved. And you think containers are going to be a good thing? Or just how, what are you thinking about that? There are two things. First, I would preface with any question, question like this with one standard answer. A lot of that has to do with execution because no matter how promising a technology is, if you don't do enough about it, then you are not going to take a key advantage of it. We're at an Oracle world. Actually, some micro system, in some ways, had this container technology many, many years ago, but they didn't take this to the next level, right? So just because container has the potential, that doesn't mean it will be there. So that's number one. Number two, can container actually get to the next level? I think it has a great chance. It does have a chance to disrupt the industry, disrupt the virtualization world, disrupt the world that people look at everything like a VN. Now people would go back and maybe I should look at everything like an application. And also you can do, especially do more. So it's all about the momentum. It kind of flattens out and there's no adoption. It needs critical mass, I guess what you're saying, right? Yes, you need to get over the hump. Containers have been around, wrappers have been around for a long time, gateways we don't want to call it. Okay, great. So final question to wrap I'll give you the final word here. What's the most exciting thing happening in Oracle Open World from your standpoint that validates some of the things that you've been working on? I think the most exciting thing is this interview because this interview is at the same time as Larry's keynote speech. The fact I'm doing this at the same time as Larry, that means you are talking about multi-cloud, inter-cloud, container. That means there is so much real attention and a real traction with this. That's it. And you've got 1400 concurrent viewers right now. So actually, I'd be curious to find out how many concurrent viewers Larry Ellison has. So it might be that you're pulling a bigger audience than Larry Ellison. We'll see. So we'll see. How is you? Thanks for coming on theCUBE. Great to hear it again. We love to geek out. Networking is really important. The applications are in charge. Computer science and some of the networking engineering stuff is really, really awesome right now. A lot more innovation going on. Thanks for sharing. And again, great to bring up Steve Herrod, your mentor and good friend of ours and now a venture capitalist. So we'll see how that turns out. He should do well. Good luck. Good luck, Steve. Good luck, Steve. We miss you. Come back on theCUBE anytime. We are live in San Francisco. We'll be right back after this short break. This is theCUBE. I'm John with Jeff Frick.