 Live from Boston, Massachusetts. Extracting the signal from the noise, it's theCUBE. Covering Red Hat Summit 2015. Brought to you by Red Hat. Now your hosts, Dave Vellante and Stu Miniman. Welcome back to Boston everybody. This is the Red Hat Summit. We're here live at the Heinz Convention Center in the back bay of Boston. Stu Miniman myself, we're happy to be home for a change. We're usually in Las Vegas or California or Orlando or somewhere else or Vancouver in your case or OpenStack Summit. Derek Collison is here, he's the CEO of AppSera. It's cloud week, past week. Derek, great to see you. Welcome back to theCUBE. Well thank you for having me, I appreciate it. So Red Hat Summit, we were just talking off camera about 5,000 or so people here. Passionate crowd, the open source dogma lives. Give us the update on AppSera. What's going on with you guys? Well, I mean, AppSera was founded about three years ago. We just had our three year anniversary and we were really trying to figure out where all this is going, right? And we have these massive disruptions around Paz and what Paz was kind of defined as early on and it's changing into. We obviously have Docker and the open container project which we were really happy to be a founding member of. And we have, you know, microservices and cloud and private clouds and OpenStack. And so there's just a tremendous amount of technology and direction and enterprises are, I think, you know, really hungry for information about what do we do with all this stuff? You know what I mean? Do we do vSphere or OpenStack or OpenShift or Cloud Foundry or is there a new thing called a data center operating system like Mesosphere or AppSera's hybrid cloud operating system? And I think what's fascinating about the technology curves that we're in is that they are shortening. And so if you looked at virtualization, it had a good 10 plus year run and now all of a sudden containers are kind of the hot thing. And I'm telling you, I think containers will probably live and be replaced within two to three years max. Everything's being compressed and by something better, faster, cheaper, right? Pick two. But it's interesting to see and it's a fun time to be part of, you know, the industry. Yeah, the whole, you know, past discussion we were talking off camera, sort of, I called it, you know, SOA 2.0. I'll refer to it sometimes as infrastructure service. Plus, SAS minus, and now all of a sudden everybody's sort of coming hard at it. What's your take on what's happening in that past world and what's different that you guys bring to the table? Yeah, I mean, I was fortunate enough when, you know, Heroku kind of pivoted from an editor to kind of one of the first passes and Google App Engine was being developed as when I was at Google and then kind of architecting and building at least the original version of Cloud Foundry. It was about taking developers and caring about what they care about which was their apps and services and say take that abstraction and leave them there. Don't make them, you know, force them into the weeds around machines and virtual machines and memory and CPUs. And so the original push on Paz was meet the developer where they are, have their application, hand it to a system and it figures out what to do with it to run it. And I think at the time that was exactly what was needed. But again, remember that compression thing. That quickly changed. And it changed for two reasons. One is the decomposition of systems, as you said, into microservices or SOA2O for the ones with enough gray hair like me to know about that. Means orchestration now is a really big deal and who can access what? Not just can I access you and find you and connect to you but am I allowed to actually do that? So orchestration and in governance of all of this across hybrid is where AppSera focuses and we've de-emphasized for our perspective even three years ago the notion of deployment. So we went from being very opinionated with at least the design I had put together with Cloud Foundry to being you do whatever the heck you want. If you want to deploy a VM, a Docker image, an app it doesn't matter. The value prop then is how do we actually orchestrate that? Govern it over a hybrid infrastructure. So Derek, it's interesting your description of really enabling the developers is something I heard real strong at DockerCon. Docker talked about how Linux containers really really deaf friendly and then they helped really focus on that. I tell you, I know AppSera was there and you're partnering, you're part of OCP. There are some in the container ecosystem that said, well, if Docker really goes where we think it is maybe I don't need Paz anymore. So can you just lay out for us? You know, how do these things mesh together? As Paz is defined as it is and what Docker is thinking about doing I think they're right. So there is no need for Paz and actually depending on the technologies that some of it's here at the Red Hat Conference I would imagine, there may not be a need for a proper IAS. You're running lightweight containers on bare metal with a policy and orchestration engine on top of it. And Docker's popularity, again, my opinion but their popularity skyrocketed because they were the new tar ball format for complex workloads. But now they're starting to go up the chain which makes sense and they should into the ecosystem of orchestration, governance, security and so their aspirations from what I can tell is is they want to be the next VMware, right? They want to replace what VMware is about. And it's very interesting when you look at that and what's going on there and then you look at the lightweight OSes like Project Atomic here at Red Hat and then Ubuntu has one and Intel has one and everybody has one. And then of course you look at IAS and OpenStack and then the Paz, it is a distinct possibility that they could get left behind. So this notion of IAS, this is a bunch of bare metal you were telling me that a lot of your work is on-prem to on-prem, private to private. Maybe talk about what you're seeing in the so-called hybrid cloud world. Yeah, well we were talking off camera a little bit about the definition of hybrid, right? And I was saying most people think it's private to public, not cloud bursting but actually really private to public. But what we've seen actually more is public to public, which we talked about meaning, well I've got all my eggs in one basket and it's not at least that we talked about, it's not as much about cost anymore, it's about I really want that service for this one app. I want Watson to be able to power this one app but I have all the other stuff let's say in Amazon, right? And so how do I do that? I don't want to have to spin up a whole new crew or retrain them on how to manage another public cloud and how to deploy and port apps to it. And so there's an interesting thing of AppSeras technology and others as well being able to smooth that over. So it's trivial, no friction to move. What you and I were talking specifically about was the exact same thing but private to private, which was how do I move workloads from VMware to OpenStack? And obviously we helped with that but we talked about another company off camera about there is a market now for hey, can we make it easier for companies who are going to move anyway to make it an easier transition for the northbound interface, the applications and the developers themselves to say, hey I can deploy and yesterday deployed on VMware and now I'm deploying on OpenStack. And the big motivation there is cost. I mean, VMware is now hopped on the OpenStack bandwagon but VMware basically OpenStack came about because it was sort of a, well I guess in part a Hail Mary against Amazon but it's really VMware, but the VTACs is really. It was a definite reaction to VMware, right? I mean, and the founders of OpenStack and I'm nice enough to know some of them and I've talked to them and they just didn't have the CapEx budget at NASA to figure out how to do what they wanted to do. Now it's interesting though, we've talked to some customers and OpenStack is definitely getting there, right? I mean, the market is obviously there. I think there were 6,000 people at Vancouver, there was 5,000 people in Paris at the OpenStack conferences. But it's interesting if you actually talk to customers they're trading CapEx for OpEx but when you do the math it's not as much savings as people think, right? You get rid of the CapEx bill and you're like, oh, I just saved $20 million or whatever but then you're like, well, how many people you got on the OpenStack project? Well, we got 150, it's like, oh, okay. So it's interesting how much money's being spent to transition from one to the other, move from CapEx to OpEx. I think the real savings will come in future years as the technology keeps getting better, keeps getting more polished and you don't need so many headcount as kind of some of the big pioneers, the big companies that have been on OpenStack. And it's enabling real business and the ROI gets better but I mean, you see the same thing in public cloud. I mean, you're going to spend more on public cloud over a 10-year period than you will if you buy on prem. Probably at least a third more, I would think, but people do it anyway because it's driving speed, agility, business process. Yeah, the thinking there has always been that if you have spiky workloads that spin up and spin down, that the OpEx spend is better and it actually works better. But people have actually said, no, if it's a stable workload it might make more sense on premise. It all really depends on one thing which is the only thing in technology that's getting more expensive is people. And so if you want to bring it in house but you have to actually use a tremendous amount of people and back when Google started, I mean, the state of the art and data centers was here. It really was. Now if you look at the sophistication from Google to Amazon to Facebook, what they actually do from breaking ground all the way to firing the data center up, it's very hard at scale to reproduce those economies of scale. And so early on in cloud days, I would say, unless it's a spiky workload, yeah, it's always going to be more money. I think the math is a little bit fuzzier now. I think it's getting close to a break even and the agility gives you the trump card just because the sophistication of these data centers and it's so high, it's hard to imagine. And taking people out of the equation. You know, that's actually a really good point is that you're right. I mean, just do we talk about this all the time where the web scale guys initially would put engineering dollars and time in to save money and the enterprise would spend on a box to save on labor. Yeah, and actually point Dave, because what are those guys focused their time on? It's they're building new applications that help drive the requirements of scale, distributed architecture that they need, and then I just have to have the infrastructure to support that. So, Derek, I'm afraid my question is, where are we with the companies you're seeing into that migration to newer applications? I mean, we were talking in our intro this morning, the average enterprise applications, more than a decade old, it's probably 15 years old. And that's the big challenge, what Paz has been trying to help people to get. So, what are you seeing? What's the update? Are you encouraged by what you've seen in the last couple of years? Is the enterprise still too slow to adopt? I think the enterprises are actually getting a lot faster. Most people right now have their heads wrapped around, oh, we have to do all cloud native apps. And it's not because that's the best thing, although it is good for certain class of applications. It's more about these new style infrastructures haven't really figured out persistent storage, like the Docker right needs one and others. And Paz has a lot of times do that are built into the platform that aren't centralized based on centralized storage. Once that kind of becomes unlocked, I think you won't see everyone saying, we have to build cloud native or 12 factor apps. And so I think we're going there. What we've seen a lot of with the newer technologies is love it for dev test, don't want to use it for production. And there's lots of different reasons. One is they think the risk is too high, they don't trust it yet. Even cloud right now for, I was talking to you guys about the red eye flight and having three meetings already. And a lot I'm talking about, yeah, we have a strategy for cloud, but not yet, right? We're still not comfortable with the governance and the compliance and the security stance around RIP, our data to know and trust it. And Absera is trying to help that and actually get those customers into the cloud, but we've heard that fairly consistency. So your bet is multi-vendor, multi-hype advisor, multi-cloud world, diversity. That's probably a pretty good bet. So maybe talk about what some of your customers are doing and what's the ecosystem that you're participating in. And just tell us a little bit more detail about how you guys are going to market and making money. Sure, so our biggest value prop is, is that we put policy governance security first, which a lot of the newer technologies don't, right? They do it after, they bolt it on, you know, it sounds interesting when I talk about deploy, orchestrate, govern, right? And we kind of reversed it, but if you look at a lot of the techs out there, pick Docker, deploy first, right? The new tar ball format, now they're doing an orchestration and then they're going to, you know, they started actually this year as well with trust, right? Around registry trust and governance. Let's make an enterprise ready. And so Absera tried to flip that on its head and say, we're going to bake in policy, governance, compliance security and trust at the core and to make it transparent. I mean, you don't have to do anything different with your workloads and then apply that to a hybrid infrastructure, where you don't have to train more people, hire more people to say, well, if I run it on your platform, does that mean I can run it on Amazon? Yes, Google, yes, software, yes, Azure, later this year. OpenStack, yes, vSphere, yes, Bare Metal, later this year. So now all of a sudden it's interesting that you can take a format that has nothing to do with Absera, like Docker, deploy it onto our platform, which we believe is one of the most trustworthy and secure in the industry. And you immediately have access to deploy in less than a second to every region in Amazon, Google, software, you know, in Azure and hardware. And that's an architectural enablement, right? That's the decision you guys made early on to bake that stuff. I mean, a lot of people would say you can't have really security bolted on as an example. Right, well, you can, it just doesn't work very well. And it doesn't scale. Comes very complex, very brittle. And so what's interesting for our customers is they come to the conclusion, along with us, as we talk to them, that what they want is a system that can accept change on both sides of the fence. So what I mean by that is, everyone talks about agility on one side, meaning faster updates of apps and new features and rolling out new versions and XYZ. But they don't talk about the rate of change in the policy and compliance side, which is usually a bad and nasty word. But if you look at a lot of the verticals, the change rate on that side is actually faster than change rate on this side. Yet the systems are designed to kind of calcify and say, I call it the default answer of no. Can I deploy my new app? No, can I get a port open on the firewall? No, can I do this? No, no, no. And so we've tried to build a system where policy was a first class citizen, so both can say yes. Hey, right, if you look at financial services, healthcare, government. Regulations are changing way faster than they can deploy new apps. And that's why those industries tend to have, the first to have a chief data officer, which is the retail, and maybe not so much. But, so, given that the web scale guys are sort of harbingers for what's happening in the enterprise, you look back five, six, seven years, and all this stuff about software defined, and agility, and cloud, all that stuff was happening in the web scale guys. If we have to look forward five to seven years, what do you think this world is going to look like? I mean, we talked again about Hadoop, that came through the enterprise, Spark coming through, in memory coming through the enterprise. What do you think the world's going to look like in five to seven years in the enterprise? Well, I mean, caveat that it's my opinion, and I've been wrong before. We talked a little bit about the fact that everything that I saw when I spent time at Google was about eight years removed to hitting the enterprise. And we're talking mostly about big data, but now we're seeing, with Docker, the orchestration engines around Kubernetes, which is an API that fronts Borg and Omega, and Mesosphere coming out of Twitter. Netflix is also becoming someone that's kind of established a way of doing things that's bleeding its way into the enterprise through their open source efforts, which is amazing. So I think, to me, you're going to see a reduction on centralized storage, a reduction on complex network topologies, right? If you can actually secure a system, you don't need VLANs after VLANs to figure out how to protect username and passwords, right? There should be a better way to do it. So I think you're going to see those things decrease because what they're built to protect, a lot of the trends coming out of Google and coming out of Netflix and coming out of Twitter don't rely on that, right? So even virtualization, right? We had a gentleman from the Kubernetes team from Google at our offices last night talking about Google runs everything in containers. They've never run on hypervisors. And now all of a sudden, and that's from day one, and now all of a sudden we're looking at this wave of transitioning from certain workloads, from hypervisors to containers. And we're not going to stop there. We're going to keep going. We're looking at transitioning from, oh, we need a big centralized storage and a database to things like Cassandra, which can run on ephemeral disk. You just spin up enough so that if you lose some, it repairs itself. Even Twitter, I think, released something where you can do it with MySQL. They don't need centralized storage. They'll spin up a whole cluster of them and you can lose a certain number of them and they keep repairing themselves. So as those trends keep hitting, you're going to have a very different stance on what are requirements of an ISR, especially around complex network topologies, complex storage requirements. They're going to go away. It'll be flat layer three IP and just ephemeral disk and it just kind of works. Then you're going to start seeing patterns change even more so beyond the cloud native apps to kind of address these types of storage, at least in my opinion. Well, it seems like it's sort of all started because of the data, right? Ship the code to the data, not the data to the code. And data has gravity and it keeps getting bigger and bigger and bigger. And that's what a challenge that a lot of your customers must have. They don't want to move the data, but they want to switch environments or they want to do test dev here and then move it over here. How do you deal with that sort of data gravity problem? How are your customers dealing with that? Yeah, there's some interesting things there because data obviously does have gravity. I worked with Dave McCrory of VMware when he kind of coined the term. I think he likes to get credit for that. And code usually is a lighter way to move around. But what's interesting is, is that if you look at one of the largest data sets that you access on a regular basis, you access it on your phone, Google. So when you type in a search, one word flies by and maybe a couple kilobytes comes back and yet it's probably touching exabytes of data, maybe upwards of 100, 200,000 servers. And so it depends on what the service is doing and what the entropy of the communications is. So for example, if I wanted to run an app on AppSera's platform and connect it to S3 and Watson, our system would say, put it on Amazon. Because it's not data gravity, it's actually the data bandwidth is what needs to suck you towards that. Because Watson's very similar, I send in a little query and I get something back. Now if the thing that's running is actually pumping data into Watson, well then maybe we want to run it in soft layer. And so we save the transmission cost. So it's not as simple, but you're right. How do you actually rationalize this world? And with IoT, which is the new cloud buzzword I guess, we don't know what it means. There is going to be a tremendous amount of data that's going to need to be processed, but it's not a plumbing problem like we had back in the RFID days. It's a identity and trust problem. It's how do you trust, not only the data, but all of the services that are acting upon it on your behalf. And there's going to be some interesting challenges to unlock the IoT. I mean, it's on its own path, trust me it's going along. But there's going to be some really interesting next couple of years around the tech to establish that trust anchor. You put Linux on an ARM processor and interesting things going to happen. Yeah, I mean I think I saw a tech where a smaller piece of technology on a chip, smaller your fingertips running Linux and on ARM and all this other stuff and it's generating data and it wakes itself up and it can live for like a year on a battery and those type of things are going to continue to happen. And right now the IoT phase is very interesting because we're generating a ton of data which most of it doesn't go to any use, kind of like big data like 90 some percent of the data never gets used. The data that's getting used is usually on a read only. So meaning that we look at it and we draw some insights. What's really going to be interesting on the IoT phase is when they try to close the loop and make it a control loop. So when the jet engine has an application that has to be trusted running on the jet engine, sending back smaller signals that then Watson let's say or something that GE is doing says we need to change something and send the signal back saying change of parameter in the engine. We're probably a decade away from that. Without a human, you mean. Right, because there's so many trust anchors to allow you to change it. And so, fascinating stuff. So Derek, you've lived through this wave of paths that we've been seeing over the last few years. I wonder if you can give us your viewpoint on how the landscape looks today. I mean, you're positioning, we're here at Red Hat. They've got a big pusher on OpenShift. Of course, Cloud Foundry and all the different pieces from IBM HP and Pivotal. And Oracle made an announcement this week. So just give us your view of the state of where we are with Paz and some of those players. You know, I think it's been a validated technology. That being said, the notion that a Paz is all about deployment, build packs, cartridges and stuff, has gone by the wayside. I think, from my perspective, what Red Hat and OpenShift have done well is they've identified this and adapted faster than the rest of the market. They merely figured out Docker's a big thing. We got to figure out what we're going to do with Docker. And if containers are a big thing, we've got to really invest in a lightweight host OS, like with Atomic. Scheduling orchestration is a big deal, right? And now, suddenly, there's an ecosystem around an API called Kubernetes. Let's bring that in. And that doesn't really have a real implementation yet. So let's use Mesos' implementation. So they're adapting very quickly. So Paz is being disrupted. Some people are going to adapt and some people are kind of holding on to, oh no, orchestration is somebody else's problem, governance is somebody else's problem. And if it goes through a build pack, we're good. And that's just not the way the world is going. It's changing quickly. Now that being said, I will say that we've seen the dawn of the age of governance bodies around open source technologies and how they can be effective, which is good to see. I mean, you got to learn OpenStack was kind of one of the first big ones. Cloud Foundry has done very well with their foundation in terms of gaining a lot of industry support. But I think the technology landscape is changing very fast. And it's going to be who can catch up, transitioning from DOS to Windows 95 to NT to Linux type of a thing. We're not going to stand still. Do you think we're going to see some consolidation? Oracle comes to market. Red has been doing their thing. Cloud Foundry momentum, we're going to have a world of dozen passes out there, or will we see some stabilization over the next couple of years? I think the market has already stabilized. I think it's Cloud Foundry and OpenShift. I mean, Staccato is a Cloud Foundry offering that does very well. Apprentice still exists in New York and Boston a little bit, space around dot network loads. But I think it's a two horse race. The problem is that the race track has shifted. So I think, if people don't realize that the race just moved way over there, they're going to keep running down an aisle that's not going anywhere. And that's being defined by Docker. So when you look at all the new passes, which there is going to be a massive consolidation around, it's mostly around orchestration of containers, Docker containers, like days and rancher and all these other ones. Days already got pulled into Engine Yard. Ranchers come out and just got an $8 million round. I think I might be wrong on that. So I think you'll see most of the small passes either be homegrown, meaning I've got an IS and I've got the GUI that I care about. Somebody helped me glue in between the stuff that nobody wants to touch, orchestration, governance and hybrid capabilities. The past world was defined in 2011, 2012. I think it's essentially a two horse race, but they need to understand that the landscape is radically shifting and they might be winning a war that nobody cares about anymore. Well, you got the big whale pass, like I said infrastructure is a service plus, you got Amazon and you got Salesforce and now Oracle jumps in and basically is making it about the red stack. I've said and I hate to say it, bite my tongue, but the rich keep getting richer in this business. Do you disagree with that? Do you see that changing? You know, I think it's fascinating if you look at the way people usually tend to react to Oracle, yet they're still customers and Oracle's still innovating and they really are. I mean that sometimes they're slow and sometimes they kind of push back, but then once they flip the bit, they have a huge customer base, they can sell into them, they do vertical stacks all the way through hardware. We talked about that a little off camera, which I think is going to be coming back into favor. I really, everyone keeps saying, oh, it's all about the software and I agree, but hardware matters too and if you can integrate those, I mean look at, I mean remember, a decade ago, well, more than a decade ago, we were all running IBM laptops with windows or some weird stuff and now everyone's got MacBooks. Why? And integrated iPhones. And integrated iPhones, right? It's hardware, software all coming together and so I think there's some interesting stuff with the big players. Most of those, I don't know if they're fully general purpose, Paz, although Heroku is, you know as a sidekick on Salesforce. Most of them what I call ecosystem expansion. I'm a SaaS provider and I have a web presence in some of your data and I have some services. You know, if you take Salesforce, now I also have a scripting language. Okay, that's not enough and eventually you get to the end where if there's enough value in that data and services, people want to write full fledged apps. But, you know, Stu could probably do it, but you don't want me trying to deploy my app on Amazon to consume my Salesforce data and everyone starts to realize that and then they say we have to curate this ecosystem and give them a safety net to land these apps. And that's Salesforce and that's, you know, several others in Oracle's doing the same thing, which is they're going to be purposely built to house apps that consume Oracle technologies on Oracle hardware and there's nothing wrong with that, right? It's going to be a good experience, hopefully for the existing Oracle customers. Where are they going to go with Docker and containers and a lot of the new tech that just six months ago, I heard people say, yeah, we're playing with it. Now, literally all three meetings this morning, yes, we have a strategy with it. We don't know when we can get it to production because there's a lot of things missing, but no, we're doubling down on that. Well, we're going to have to put it in their portfolio. That's no question, how aggressively they adopt it. We're out of time, Derek, but I wanted to give you a last word around Efsera. So business update, you know, where you're at, you're raising money, selling the company, buying companies. Yeah, no, we did a majority investment deal with Ericsson about eight months ago. We were 21 people at the time, we're now at 81. So we're hiring like crazy. Ericsson has funded us to a point where I don't have to be on Sandhill Road anymore. And that was kind of the opportunity that I thought we needed to take. As you see with Docker, you know, in terms of the investments they got, you're going to need $100 plus million of operating capital to take advantage of some of these trends. And I wasn't going to be able to do it with traditional VC methods. So we did that, it's been an amazing relationship. They've come out with HDS 8000, a partnership with Intel that we're a key part of, which is going back to marrying hardware and the software and the services altogether. That's great. We're onboarding the first 10 customers, right? Some of the things we have approached and done are very hard. It's taken us a long time to get them right, but they're rock solid now. And we're getting a lot of great feedback from early customers. And we're not a pass, right? We're a hybrid cloud, kind of an operating system type of a technology. And so we co-exist with the world and we co-exist with Docker and, you know, 2015 and then into 2016 is going to be a great year for us. Helping you simplify and manage all that diverse complexity. Derek Carlson from AppSara. Thanks very much for coming on theCUBE. Thank you. I appreciate it. I appreciate it. All right, keep it right there. Everybody, Stu and I will be back at Red Hat Summit from Boston. After this, this is theCUBE. We're live, right back.