 From San Francisco, it's theCUBE, covering Radio 2018, brought to you by VMware. Welcome back everyone. We're here with theCUBE in San Francisco for exclusive coverage for VMware's Radio 2018. I'm John Furrier, your host. This is the event where everyone comes together in the R&D and the Organic Engineering Organization of VMware to flex their technical muscles, stress their minds, compete for the papers and also get to know each other. And the key person behind this is the Chief Researcher, David Tenenhaus. Thanks for joining us today. Thank you, John. I'm really glad to be here. So you're the Chief Research Officer. We're going to look at the company-wide agenda. But this event is more of a special event organically. Talk about, for the folks out there watching, what's different about this event that goes outside the scope of the top-down research? Yeah, this is really for the developers, by the developers. So when you said I'm in charge, I'm definitely not in charge. And we have a program committee. There's a program committee chair. It's much like the way an academic conference might be organized, where there's kind of a group of academics that watch over the content. In this case, we have many hundreds of folks that submit proposals into radio. They can't all get selected. It's very competitive, because in addition, if you get accepted, you get a ticket to radio. You get to attend. So everybody really wants to do that. Talk about the organic nature, because this is one of the things that I've seen that's been part of a world-class organization, like Amazon has their own process for it called the Big Idea. They have certain working documents, the process to foster any idea across the organization. How important is that as part of the radio? I mean, literally, it's anyone, right? Well, it's not just radio. It's important to the whole company. So I think of this as when you're working on innovation, you're going to have sort of a breadth component. You want everybody doing a little, and some of that's going to be incremental. One thing I learned in a prior role at a different company is if you add up a lot of 2%ers, that's how you can double things and keep on Moore's Law every year. So you're going to get some of that, and you're going to get some really disruptive ideas. So from a top-down point of view, we try to drive some disruptions. Some disruptions show up organically from the troops, and a ton of that breadth stuff shows up. I'm honored to be here. It's the 14th year, and some t-shirts commemorating the key milestones from way back in the day. This is the first year Press was allowed in. I noticed a handful of folks came in to kind of document this. A lot of the brightest minds in VMware here. Again, great to have us. We're super excited. But share with us what's happened over the years. Give some examples of where people were coming together, where there's a collision of ideas, and just that combustion that happens. Can you share some stories around the key notable, or potentially, as Raghu pointed out, there's been some misses, too. Yeah, you're going to get some of that. I mean, you've got to take risks. Not everything's going to work. And just to speak to misses, what I've learned in the innovation and research space is as much as anything, it's about timing. You realize, it's pretty rare that you completely technically miss. Usually engineers have an idea. They'll figure out a way to make it happen. Then the question is, is this at the right time? Are the customers ready? Is the market ready to go in that direction? So that's just to talk to that. Timing is a big deal. Well, there's never a miss, too, in R&D, because if you, like Pat Gelsen said, understand when it's got to be recast it. Know if it works or not. Did it mean it worked? Just understanding. Those are the ones that actually, I feel, what I really hate is if, for some reason, we have to end a project and we haven't actually gotten to the bottom of it, right? And so you don't know yes or no. And sometimes that can be the kind of time to run out, right? You've decided, well, even if it works, it's too late. But getting back to some of the examples, I'll focus on some more recent ones. We had some really interesting work come together on containers. And there were some folks that, and this is going back like four years ago, containers aren't a new story and certainly not for VMware. But around four years ago, there was a proposal, a radio, that had to do with, let's make containers first class citizen on VMware's platform, okay? To the top level, that makes great sense. Let's go do it. Containers are great for developers. The IT folks still want the isolation they get from VMs. Let's put these together really effectively. So that was top level. There was a next level of the idea that said, gee, at radio a couple years before, there'd been this idea of being able to do something called VM fork or being able to clone a VM. And saying, and this came out of our end user computing group, the VDI folks. And if you think about it, if you've got a virtualized PC, you want to be able to clone that so you can start these up really fast. And the container folks said, hey, we've got the same problem. Could we actually try to make use of that technology and use that as part of our bigger container push? So those are examples of things that came together at radio. And they're also examples of things where the market timing may not have quite been there. So we went out with the container work. That was actually post-radio. It was funded. We incubated it. You've got vSphere integrated containers. Hit the market exactly the right time. Timing right there. Timing right there. But what we learned as we actually started doing trials with customers was that they didn't actually need the instant clone on the containers. What they needed is throughput. They wanted to know that they could do large numbers per second as opposed to you'll get that container really quickly. So as the team went along, they actually shifted away from that fork idea. We'll probably come back to it when the time's right for it. Well, you have a nice little positioning there. I like the timing because, by the way, entrepreneurial timing is the same way. You go outside. I was in the VC. Okay, so you know. Okay. Timing's everything. Is the timing is that? I mean, sometimes you see that entrepreneur wicked early on that going, and they keep scratching that edge and finally they get it right. Yeah. The art of the timing. But also the art of knowing when to, what to keep an inventory. Pat mentioned the VCloud Air as an interesting example. Yeah. Recognizing an abandonment there. Okay, hey, let's just stop, take pause. Let's use what we have. Do something else. Do something else. You got to do something else. Yeah. And by the way, along the way, in parallel with VCloud Air, we had built up these VCloud partners. And that's phenomenal, right? So we have, you know, people think in terms of the couple of very large public clouds, but we've got literally thousands of people running public clouds in either specialized markets or particular countries that are running on our platform. And, you know, that whole VCloud Air effort helped push that forward. So where were you at VC? Just curious. I was actually in a company that fits with, with, you know, sort of my role in research and innovation. I was in a specialized firm, boutique firm, New Venture Partners that specialized in spin-outs from large companies. This goes to the timing, right? Yeah. Yeah. I'd previously been at another large company, you know, and whenever you have a research portfolio, you're going to have some projects that you started. They were technically successful. That's your first notch. Then you go look and say, hey, can I find a business model for it? Some of these are both technically successful. You find a business model, but you had anticipated that the company strategically was going to zig, the company's zagged. Now, this is a great opportunity. It doesn't quite fit. So, you know, we did those spin-outs. Well, just, I love the perspective too, you said earlier, David, around not getting to the bottom of it. And that's the most frustrating part, because you just like get some closure. You know, like, okay, this thing, we took it to the end to completion. This is not going to, good try, guys. And we know why. And you know why. Now, let's take it to the next level. Now, the market we're living in now, I heard with Ray O'Farrell, I was talking about earlier, with earlier, we talked about the confluence of these big markets coming together. Infrastructure market, which is kind of declining on paper, but cloud is filling the void. Big data is becoming AI and blockchain over the top. These are four major markets and the center of them, intersecting all these nuances, security, data, IoT. Governance. Governance is like, so there's some sticky areas that are evolving based upon these moving markets. Opportunity recognition is another one. So this is what you're kind of doing now, the research. Talk about opportunity recognition. We definitely do that. And I just, I do want to say on the infrastructure side, you know, something to recall is that as people, you know, they've got their private clouds, those are individually getting actually bigger as they consolidate. But now with IoT, you're seeing edge computing pop up, right? So, you know, the private infrastructure doesn't go away, it moves around, right? It's like a liquid. And you pour it from place to place. It sounds like Ray O'Farrell was talking about in his keynote, early days of VMware. Again, it computes the center of this. Right, compute, but, you know, I'm a networking guy. So, you know, we've grown that. And I think that, in fact, you know, more and more as we make progress with software-defined network and network virtualization. And if you think about that, so, you know, let's look at that. So, it computes definitely, you know, at the center of what happens in the data center, in the cloud, right? You're going to want to be able to, you know, string those pieces together. So, today we've got AirWatch. I think that's strategically really key because it gives us a little bit of presence on the edge devices that touch people. That's one of the ways information gets from the physical world to the virtual world is through people. It's an edge device. People are things too. IOT, right? So, we're, you know, working hard and that's one of the projects that we incubated in research and it's now become a business at VMware is to get that presence right at the edge at the gateways that bridge between the things that are connected to the physical world and bring it into the virtual world. Now, if we can put our software-defined network between all that, so you've got it between the public cloud, the private cloud, you know, the mobile devices and people's hands. And the on-premise state center. And exactly, all of it. All right, so here's a question for you. Full out, this is one of those trick questions. Is the cell phone an edge device or an IOT device? Well, I think it's in many ways both. And what I think of it is more of a gateway. If you think about the IOT world, you have the things. IOT is a strict definition, though, in your mind, right? People refer to IOT as more of a sensor thing to a physical device. I tend to think of it as it's got some connection to some physical device, right? It's able to bring information in from the physical world. Okay, so now you look at your cell phone. It can bring information. It's got that microphone. It's got that camera, right? It can bring information in. Can it do a physical person? It can put information back out. Yeah, through a physical person. Well, look, I've been in the space for a long time. Going back to my time at DARPA, we set out to create the IOT world. There wasn't an accident, right? We looked at this and said, okay, the main way information gets between these two worlds today is through human beings, right? The way I used to explain this to the generals is, you know, we can't keep putting human beings in the direct line of fire of information technology. So we've got to get these devices, got to get all these sensors. It's taken a long time. This is, you know, again, timing, but if you look at the research world, this has been a 20-year pull. By the way, incredible work you've done, by the way. From there to here. It's been amazing. Just to, you know, pull this along. But, you know, so when you look at that cell phone, it's got some of those sensors. It's got, you know, actually, a whole pile of sensors in the phones today. It's got actuation, the ability to put the information back out. It's also a gateway. Because typically, you know, particularly through its Bluetooth functionality, and as we get Bluetooth low power now, so it's also acting as a gateway to connect up other devices around your body network. Personal networking, whatever comes on your physical presence. You know, turn that around, and it says, you know, in the IOT world, we've got to manage gateways. We've got to make sure gateways stay secure. Because they're really going to be the sort of main perimeter, the line of defense. If you think about all these things that are going to be out there, as an industry, we're going to collectively try, very hard to secure all those things. But let's be realistic. They're going to be supplied from a wide variety of companies, and they're going to last longer than people might think. How much of those devices are operationally, operation technologies, non-IP, versus, not IP, Internet Protocol, Internet Protocol. Yeah, yeah, non-Internet, you had it right. The VC in the brain there, the VC, IP. I'm like, get that IP right. So, Internet Protocol devices, which has some challenges, but that's getting fixed, versus OT, just sensors, proprietary. Yeah, well, either proprietary, let's say, you know, maybe an industry standard, but an industrial standard. So, today, a very large, you know, large fraction, particularly, you know, you asked about how we focus at VMware. Well, one of our foci is, we're about what are our enterprise customers going to need? So, when we think IOT, we're not really thinking that much about the consumer devices, we're thinking about those enterprise devices. So, a lot of those- That's where you watch Mike come in, so the employees still have phones, though. Employees still have phones, so that's why I said, so there's the human interface, we want to be there. And there's the other enterprise interfaces to all these sensors, that could be in a factory, it could be in a smart city, any number of places. So, as we pull information in from those, we're going to find that they come from a lot of different suppliers, and they're going to last a long time. You know, even if you buy a device, it's got a three to four year lifetime, probably 10 to 20% of those still going to be around 10 years later. Right? If you're smiling, because you know that in your home, you have some Wi-Fi connected devices that are a little older than they probably should have. Yeah, and they have full processing capability threaded processors on it. Right. Which could be running malware as we speak. Exactly. So, as I said, as an industry, we'll try to secure those really edge things, but the reality is we're going to have to draw the line at the keyboard. It's a lot more secure work. I totally hear you. I mean, the light bulb could have a full thread on there. You know, the surface area is so huge now. And there've been attacks on light bulbs. Yeah, I know. So, I got to ask you a question. So, I got to ask, because you bring out this networking edge, which I love, by the way, I love anything that's network, because I think this is the future of work. How is the future of work impacting some of the R&D you're doing? Because, to talk about air washers, having more mobility, the human impact society, whether it's mission-driven and or just human collaboration going digital, you're going to need to have policies. You need to have a networked society. Right, this is super relevant, but it brings back that future of work. It does. And so, a couple different aspects. You know, one, you know, which just relates to a point you raised, is if you look at something like our Workspace One product, I don't know if you've had a chance to do that, it's kind of a win-win, because you get one portal. So, you know, an employee for an enterprise, they've got one portal, they get access, you know, it doesn't matter whether they're getting to a web app, they're getting to a, you know, a DVI supported, you know, application, they're getting to something that's on a server, something on a SaaS player, right? They get through that portal. So, for them, it's convenient. I mean, for me, as a manager, I love this, right, because whether I'm on my cell phone, I'm on a laptop, doesn't matter. I can get to the same, you know, expense app. You don't need to carry two phones, my work phone. And I can do all these approvals really easily, right? So, I also don't worry, you know, I don't see the difference between which device I'm on. At the same time, that you're delivering that convenience to the user, you're delivering governance, because the IT team can be deciding how that portal's populated, how things are connected, right, and how the wiring works. All the authorization, you've got a common identification system and all of that. So, that's kind of, you know, very specific to, you know, let's say, near-term changing the user interface. In terms of the broader future of work, clearly machine learning is the big story here, right? And I think that what we're going to see is, particularly, again, an enterprise, more and more need for data analysts to be able to look at the big data. We're going to see sort of more and more use the machine learning technologies that's going to, you know, basically creep in everywhere. And we're getting this at just the right time. So, if you want to think about future work in the big national and international scale, what you really sort of stop to look at is say, gee, okay, these machines are going to do all this work. What about the people? And, you know, a lot of people then, therefore, get concerned. Gee, the computers are going to take away all the jobs, right? That's you get- I think right now we're worried about fake news and real content. Well, let's come back to that one later. But there is a sense of, gee, you know, the computers will take on all the jobs. And, you know, what I think people are not doing carefully is looking at the demographics. Because if you look at basically all the developed economies for practical purposes, we actually have a demographic problem. Our problem is actually not a surplus of workers. It's going to be a shortage of workers. In fact, actually in the U.S. right now, you're starting to feel this. Now, that's at the peak of the economic cycle. So, of course, you feel it, you know, a bit. There's trained workers, too. There'll be peaks and troughs. There's also people who qualify. Right. So, I think the thing we really need to look at is how do we do a much better job at matching, you know, sort of workers and both folks coming into the workplace, people with existing skills, two available opportunities, because actually we're going to have a shortage of workers. And it's not just sort of the U.S. and Europe. I mean, Japan for a long time. China headed to a shortage of workers. I was out in Singapore not too long ago and was surprised to find out not just that they're concerned, but they went and looked at the Southeast Asian countries around them that are their markets. They're looking at a shortage of workers. So, you know, if we didn't have something like, you know, machine learning and AI coming along, we'd be sitting there saying, how are we going to keep our economies growing? We need augmentation for sure. We need this augmentation. And it's coming at just, you know, talk about timing. You know, it's coming at just the right time. Now, there definitely are going to be some tough transitions along the way, right? So, we definitely, you know, for example, as autonomous vehicles come along, we've got to figure out, okay, all those people that are driving vehicles, you know, what are they going to do going forward? But let's not kid ourselves too, you know, if you've got trucks moving around with high value cargos, you're not going to leave those unattended, right? We're going to have to figure all this out. So, there's going to be a lot of interesting opportunities. What's your take on blockchain? Well, first of all, GDPR, real quick, train wreck, useful. I think it's, you know, if you backed up and asked me, you know, four or five years ago, I've said train wreck. And largely because, you know, we still don't have this sort of, you know, kind of international consensus on what the rules should be. But you mentioned governance earlier, it certainly needs to be at the center of the action. But, you know, if we take a look now, it seems like it's showing up at just the right time, right? And, you know, in that sense. So, I think, you know, part of what's happened is over the intervening years, a lot of countries outside of Europe, because they realized these regulations would apply to them. They've worked with the European regulators to, you know, help the regulators understand the technology, you know, help the companies understand them. That's a good politically correct answer. I'll just, I think it's a shit show personally, but, you know, well, I mean, it's going to force people, it's like, why 2K in money making, but why'd you never happen? It's forcing people to really, I think the value of GDPR is, the big company's going to get hit hard in some suits. Just the trolling thing bothers me, just the trolls that could come out of the river. But I think the positive, it puts the center of the value proposition, making data not a one-off, like backup and recovery, has to be core to technical operations. And making privacy something that's really, you know, in that first class category. Yeah. You know, it's, as I said, like, it's a great first step, but, there's a big but. There is more to be done. That's hopefully they don't go after us little guys. All right, final question, Blockchain. We are super excited about Blockchain. You have team working on this. I am super excited about Blockchain. Talk about your view on Blockchain, why are you excited about it? Obviously, you know, we feel it's very efficient, makes inefficiencies efficient across all industries, your thoughts. Okay, so again, we look at things through this prism. What are, you know, enterprise customers going to be looking at? What do they want? And, you know, so we're not, you know, just, you know, I think you're in the same place. We're not looking at the cryptocurrencies, right? That's not the thing. And in fact, we're not even looking at cohabiting on the Bitcoin Blockchain, because do you really want to run your business, you know, in the same place, that a whole bunch of other people are running illegal businesses and, you know, the whole thing. And by the way, this is right. Technical issues. We'll get to that, right? We're going to get there, right? But just even as a starting point, so we pretty quickly looking even, you know, three, four years ago said, okay, enterprise is not going to want to go that way. But this idea of a federated ledger, right? So if you can make federated ledgers and we can have reusable technology, that means now, if I want to federate with other companies or other organizations or, you know, or you need companies federating with governments, or governments federating with each other, anywhere you want to pull together essentially a club for the exchange of data with a persistent record of what happened. You've now got a common way of doing it, right? Or we can drive towards that. It'll be, you know, there'll be a standardization process to get there. But so it's not, you know, to me, federated ledgers means lowering the barrier to federation. And I think that's pretty exciting. Whole bunch of places, you know, supply chain clearly won, financial technology, but... David, we got to get spent some time. You have to come in the studio. I'd love to explore some of these great topics with you. But I got to ask you one final question. You know, with your history, you're only back to ARPA, the ARPA days, and looking at really the beginning of the information superhighway, IP, connecting some universities together, to today and the waves that have gone through. We've talked about standards, the OSI stack. You had all these grandiose standard plans. Not all of them have happened exactly as planned, but de facto standards play a really important role. It galvanizes community, gives people a guiding principles, North Star, whatever metaphor you want to use. The key is the enabling, disrupting technologies. De facto standard. What's happening now in your mind that you see out there that's starting to emerge as de facto? Because certainly there's a lot of standard things going open sources, you know, for tier one citizen growing rapidly, which is greatness, cloud is booming, unlimited resources compute, fingertip compute. Oh, this is good. Yeah. What's, and all these new standards, I got Kubernetes, I got this and going on. Well, again, they're de facto. Right, Kubernetes is an interesting example of basically open source meets de facto. And that's pretty exciting, right? I mean, we're excited about it. I think people are often surprised we're a fan of open source. Yeah. And I guess what I just like to sort of back up a notch because what you touched on is de facto standards and whether it's open source or not have suddenly become a lot easier when I say suddenly over like a 10 year period. And I think what's going on there is this is part of the change to software. So if you're talking about hardware and you got screws and you got threads, these physical things have to match and they have to match exactly, right? Say when you travel overseas, you need to carry converters, physical converters to convert from one thing to another. So if you want to interoperate, if you and I want to have stuff that interoperates, we need it to build like either do the same thing or have a physical adapter. There was a cost to not having a standard. If you think about it in the software world, we can build software converters, right? So if I've got, say we've got two or three or four or even 50 de facto standards in the software world, blockchain, so there's 50 new things. Everybody launches their own. Pretty quickly, the market will drive that down to a small number and then you can put software converters in place. So we no longer actually have to get to one. Yeah. That's a software economic model. It's a big change. And that is huge. So by the way, we had Dirk Hondel on a KubeCon Love is open source mission. Just a shout out to you guys doing a great job. You guys at VMware, certainly we know, love you opener of these shows. Final prediction, final question. Give us a prediction. Give you a prediction. 2018, second half of the year. What's going to happen? What's going to be a notable thing that you see out in the horizon that might happen in the marketplace? That might be notable for people to stand up and pay attention to. I think we're going to see some significant developments in the blockchain space. And it's going to be in the category of people starting to announce real deployments. And if you're sort of looking at that timeframe, you've had a lot of different enterprises try things. We've had people kind of dabble at things. I think you're going to start seeing some people really move significantly in that space. And you think, like another first follow-up on that, you think like in the database world now, where by the way, it's okay to have a zillion databases now, because you talk about databases and you have a distraction. But it can solid data down to a few layers. It's okay to have a view Friday of blockchains. I mean, there's no one block chain. Correct. So that's where I think, as I said, you're going to see actually a bunch of these deployments. They'll be using different technologies. And then the fun race starts, right? As people consolidate, especially with open source, they swap ideas, we boil it down to what's the best of the best. We've got stuff we're doing, certainly, to knock the throughput down. Sorry, throughput up, latency down. And we think we've got a very scalable approach. And most important, it's something that's real. I don't know if you've talked to people about our sustainability. It's key value for VMware. A lot of great centers. You can imagine, we looked at blockchain, we looked at proof of work, and we said, that's proof of energy wasted. We're not going there, right? Gotta make it more efficient. I think you're going to see more and more folks focusing on things like Byzantine, fall tolerant, ours is scalable, SBFT. Performance is key. And the energy is a huge problem. But performance and acceptable energy. You can't just waste, it's immoral to just waste energy. And it really goes against what a lot of the whole IT industry's built up. I think over the decades, we've done a lot of things for the goodness of society, and we've got to stay the mission. I think as the more, I won't say mature, but big world-class organizations join in, I think that'll straighten itself out. And certainly, as any evolution would see the web, I remember dial-up in AOL. It can't go as fast as this mini computer. Well, you don't get it. It's the web, okay. David, thanks so much for coming on. Appreciate it. Great conversation. Here at Radio 2018, I'm John Furrier. Cube coverage of VMware's annual 14th year conference, Radio 2018. Thanks for watching.