 Hi, I'm Stu Miniman and welcome to the Wikibon Studio. Join with me for this segment is Brian Gracely. Here in the middle of the summer, 2015, and want to really just do a rapid fire run through some of the hottest technologies in technology, maybe splash a little water on some and understand where we are with the maturity of where things are going. So Brian, let's walk up the stack a little bit. Start with one that we've been looking at for many years, substrate, we at Wikibon have said transforming the way that architecture is going to be built. Let's start with flash, what's your take on flash? What are you hearing out there? We're going to come to the hot technology. I got to put flash in the hot technology. We're seeing EMC is doing extremely well at Extreme I.O. Pure storage is still growing, still took a huge round of funding. Solid fire took a huge round of funding. The cost of it's coming down, deduplication's helping it. I think people understand it, I think they trust it. Has it taken over disk, not yet, but I think it's, I'm going to put it in the hot category. Yeah, and Brian, actually, I want to poke at something you said. So you talked a lot about the all flash arrays and I think most people tend to try to think of a technology in a homogenous, it's like one bucket. When we first got in, I remember 2011, David Floyer wrote a piece and he said we've got kind of server based flash, you've got all flash arrays and then you've got what's going in the hybrid or just adding to those. We still think there's a ton of innovation to happen in flash, even if we're now about six years into it, if this was a 10 year wave, we think there'll be more innovation in the next four years than we saw in the last six. The server based or leveraging really high speed, what David Floyer calls flash as memory extension is kind of the next early wave that's coming in flash. All flash arrays and hybrid arrays, more mature, still growing at a high growth rate, probably going to see an IPO either from a pure or a solid fire in the near future. And I agree, customers are ready for flash. It's no longer kind of what I don't get it or why it's where are they doing it? How does it fit? What about cloud, how does flash fit into the cloud? So what we're seeing in the cloud is we're seeing more and more all SSD offering. So people, whether it's Amazon, whether it's Azure, offering all SSD storage for persistent storage. So they're basically realizing that people want IO, they want predictable IO, they want it in the cloud. So we're seeing that more so. The way that's deployed typically is going to be what you call server sand or flash inside of the server itself. But we're seeing that trend happen just as fast. And in fact, the prices on that are coming down and I think that's helping to drive the overall pricing in the market down. Okay. What about any red flags or for issues with flash that we should cover before we cover the hyperconverged service sand piece? The only red flag to me right now is I think some of the vendors are getting a little over their ski tips talking about the all flash data center, only for the reason that if you look at how much capacity in the fabs to build all flash, I don't think it's there yet. So if somebody flipped a switch tomorrow, not enough capacity to do all flash data centers, but it sounds like it's a great title. It's a great buzzword. Yeah, and absolutely. So the nuance I think we'd give from Wikibon is if I'm doing all flash for really my performance, my active data, that's where flash is going to be. I'm still going to need capacity. We actually coined a term called Flap, flash plus tape because you're right. I'm not necessarily going to put everything that I need to archive on flash. That doesn't make sense. There's still room for capacity disk. There's room for tape. Those technologies aren't dead. They just have their use cases. They're cash cows for many of those companies and that sort of flash going for it. All right, let's move. You brought up, if we look at the real web companies out there and how they're leveraging storage, most of them aren't leveraging storage arrays. They have flash and other storage technologies built into the compute. What we at Wikibon dubbed server sand. So Brian, you and I both have a lot of background in kind of the convergence, how we're trying to simplify the compute network and storage of infrastructure. Where are we in your standpoint of converge and server sand hyperconverge being the newer piece? Yeah, so I break it down into two things. Number one is there's a lot of different terminology. Converged infrastructure, server sand, hyperconverge. To me, the one big trend that is absolutely happening is more and more customers are asking to buy things packaged together. And we're seeing this in virtualized environments. We're starting to see it even in containerized environments. But that trend of packaging things together, whether it's converged or hyperconverged, whether that's a V block or what Nutanix does or what Evo Rail is beginning, all those sort of things. Like that trend I think is happening because customers just go, it's easier to deploy and easier to deal with. Now, in terms of hyperconverged, that's where things start, we're still in that stage where people call it different things to sort of manipulate what it is. What's your take on that part of the market? Yeah, great point. That's why we put our vision for server sand because it's really, really early in this market. So even in the last two years, we've seen some maturation. So most of the solutions today are what we would call really an appliance, which is, I've got my compute, it's got the storage baked into it and a software layer that helps manage all of it. You've got some software-only solutions, things that are like hyperconverged. So if you take Nutanix as the leader in the appliance market, you've got companies like Maxta, who says, take my software and make pieces of it. There's a company EMC bought called Scale.io, which looks kind of like hyperconverged, but the way it is deployed is typically, it's hundreds or thousands of nodes, not as much functionality, and hits very different use case. It hits more really the service providers we talk about. If you talk to some service providers or even the web scale guys, they don't even think of it from a storage standpoint, they just look architecturally. An example I've used often is Facebook, when they do talk about it open compute project, they have five configurations of server that they pick, and that has the storage built into it. It's not like they say, okay, I need my storage tier, but if I'm database, I have this type of server. If a mobile app, this is kind of what I need, whether it's optimizing for performance, optimizing for latency, optimizing for capacity, and then I build those at massive scale. So we're really early in kind of this hyperconverged market, and we think there's plenty of room for maturity and expansion, and we're pretty well known for saying that within the next 10 years, we think that this new architecture, which is very much software-based, built for distributed architectures, is gonna decimate a lot of the traditional SAN and NAS marketplace with the external storage array. Yeah, and that'll be a big thing to watch for your sort of traditional storage companies, your NetApps, your EMCs, HP to a certain extent and some others, and they're already beginning to see that trend happen. Any big red flags you're hearing from end users or just architecturally about hyperconverged or converged? Yes, so traditionally when we built an environment, I have to worry about my application and how everything works. So we're still kind of knocking down application by application, understanding it. For the mid-range customer, many of these products work very well, but even for mission critical, the kind of number two startup in the space is SimpliVity. SimpliVity starts all with multi-site environments with mission critical applications. If I took a Hitachi and IBM or EMC environment and I put in SimpliVity, the storage looks very similar for them. So red flags are, we're still building up the scalability, the maturity of the platforms, the global scale of the support on them and really understanding how all my applications fit, not just the first couple of projects or moving them on. We don't have too many customers that have thousands of applications running on one giant pool. It's typically multiple clusters were going to be set up, it's spreading out, but it's very promising even if there's some nuance that we need to work out. Right, right. All right, so I guess the other thing I'd look on that is I'm starting to talk to some of the newer players in the space as to how they fit into service providers and cloud environments. At Wikibon, we really have it split out today for Hyperscale, ServerSan and Enterprise ServerSan. How that kind of merges down the road is still a little bit foggy in the crystal ball for us. Yeah, absolutely. So I guess if we go from there, that the infrastructure, we've talked a lot about cloud in some of the various segments, but OpenStack is really won the marketing war for what private cloud's going to look like, especially if we're talking about things like we're talking about our void lock-in. We want to have some flexibility as to whose stack I choose from. Brian, what's your take on OpenStack today? So from a OpenStack perspective, OpenStack started off with big ambitions. We're going to take on Amazon in the public cloud or we're going to be the dominant private cloud. I think that strategy backfired a little bit. They really never put a dent in the public cloud. In the private cloud space, it really is. It's becoming VMware and OpenStack. And we've seen the consolidation or sort of reduction of that overall OpenStack market. We've seen either companies getting acquired, Cisco's acquired a couple of them, EMC acquired a company. Morantis is really sort of becoming the dominant OpenStack player along with Red Hat is sort of their de facto what's in the operating system. So that market's played out the way we expected. We expected it to consolidate. There was a lot of players that's consolidating. You know, to me, I think what to look for as you're looking out in OpenStack is, I think you mentioned it, it's getting boring. It's really, it was a lot of hype for a while. It's now getting boring. Hopefully that whole community begins to focus much more on operations because that's been the thorn in its side is lots of features, really no focus on operations. And so it's not getting deployed as quickly. But you know, they start to focus a little bit of that. IBM made a big acquisition. They've seen some talent move over to there. I think we're going to see them step up as a big third player in that space. But yeah, my hope for OpenStack is the hype sort of mellows out a little bit and the reality of it becomes a valid competitor with VMware. Yeah, you know, thing I'll point out, if we talk about maturity, we're only five years into OpenStack when it came from that very audacious, you know, goal of what they were doing. It's like, right, were they going to try to kill off Amazon? Obviously that didn't happen. Or the early players were all saying, hey, we're going to free you from having to pay for VMware. And that hasn't happened either. So the new role, the article I wrote about it was OpenStack's trying to help be that integration engine. So how do I adopt new technology? How do I keep my infrastructure flexible so that I can work across multiple environments, not be locked into any one component? But you know, we're talking, it's compute, storage, network, the whole management layer on top of that. You start adding things like, you know, containers and, oh wait, I want to do virtualization and bare metal. I've got like, you know, so traditional projects and you've got things like Ironic. There's so many projects out there. The red flag I had had for the last couple of years is that there were two main ones. Number one is, you know, we really have project sprawl. And what's part of OpenStack and how do we fit and how do we mature all those pieces? And OpenStack came out with really the big tent to try to help solve that. And secondly, all of the vendors that are part of this, how do I understand what's in there and what works? If you talk to the storage companies and they say, oh yeah, I fully support OpenStack and you talk to five different vendors. Some of them, oh, I'm doing testing. Some of them, I wrote the code. Some of them, you know, it's just across the board. And so we're actually going to have, you know, well understand, you know, what gets the label from the foundation as an OpenStack solution. So the comment I made at Vancouver this year is, I think I understand how the red flags can go down and I can really move towards that maturity. The thing I will point out is that the maturity of the whole solution still needs a little bit of work. I read a great article recently talking about how, you know, somebody that really lived on the operator side of the world said that the developers own OpenStack, they're not listening to the operators and therefore every single company he worked with has to fork OpenStack to actually get it to work and then that's not getting back. I don't have something that is a repeatable solution which is really, we were talking in the last piece about, you know, convergent infrastructure and, you know, hyperscale distributed architectures. I want to have a repeatable, easy to manage environment and that's not where OpenStack is yet today. Yeah, there's really two big things that I'm sort of excited about from an OpenStack perspective looking forward. Number one is there's this new emerging trend where a number of companies, so Cisco acquiring MetaCloud, IBM acquiring BlueBox, we're seeing Platform9 as a company, basically essentially people saying, you know what, even if it's too hard, we'll help manage it for you and they're doing some unique things about, you know, essentially making it a managed private OpenStack cloud. It could be on your prem, it could be in their data center and I think we're going to see with Cisco's backing and their revenues, IBM's backing their revenues, we're going to see that model take off more, which I think is a good thing, get to using the infrastructure to focus on the application. And then the other big thing, and this came out in Vancouver, we'll see it at the OpenStack Silicon Valley event, more of a focus of how our container is starting to work in. You know, OpenStack was always virtual machines, containers are starting to come in, starting to see Morantis do some things to people. So those are two very positive things in terms of people saying, how do I get from here to there and how do I just make this work? Yeah, the last thing I'll point out, there are some that have kind of griped that the big players are coming and taking too much involvement in the environment. So it's not just, you know, Rackspace and a bunch of startups who are now bought by all the bigger companies from Rackspace, they're still standalone, but they're doing other stuff even. But you've got companies like IBM and HP and they've bought some companies that people that really understand what's going on. And I think one of the things that HP and IBM are really good at is they know that they need to deploy this in thousands of customers and it needs to be a repeatable solution. Thing that I always gave HP really good credit for was if you deploy an environment, you know, when it shows up at your location, the IP addresses are all taken care of. It's pre-configured, right? Because that's what they really know how to do. There's this rack that comes in and it's not, I'm going to spend weeks to get it up and running because they need to do that at massive scale. They're leader in servers and they put a thousand people and a ton of money into doing OpenStack. So hopefully they can help with that maturity cycle so we move a little bit more from the developers to getting the operators happy and getting more users there, right? All right, you brought up containers. So, you know, Brian, you know, I feel I've been a broken record for the last year. Everybody's like, okay, Stu, what's the hottest technology that you've seen? And I, you know, my answer's always been, you know, other than Docker. So we are so new at Docker today. I mean, we talk about the maturity, you know, 24 months ago, we weren't talking about Docker. It's like, you know, the first interview we did with the Docker folks was, I mean, it was 2014 at the beginning of the year. We talked to Solomon Hikes and, you know, the company's only about two years ago. So where are we with the maturity, the reality of kind of the containers in general? Containers have been around longer than Docker but Docker is really the one that's brought it to the mainstream. So, yeah, so from a container perspective, a couple of big things. Docker's leading the industry just in terms of driving interest in it. You know, half a million, more than a half a million downloads in the, or half a million, 500 million downloads in the last year. So there's hyper-interest. It's not just interest or buzz. It's hyper-interest. The other thing, the ecosystem around it is starting to do some really interesting things. Docker announced a bunch of plugins for networking technology, storage technology, security. So that's a good thing, right? It starts to follow the model. It says it's not just one piece, but it lets the ecosystem play. And we're seeing really aggressive, interesting technology from CoreOS, HashiCorp, from Docker themselves. We're seeing some of the traditional vendors. EMC was doing some things with Docker and storage. We're seeing Cisco start to build some plugins around that. So that's a very, very early days. I mean, for anybody who's a VMware person, it's sort of like comparing it to VMware 2.0 or so forth. I still think the nice thing we're starting to see, and some folks will take this as a negative, I look at it as a positive, we're starting to see actual production use cases come out where people are saying, it ain't easy. It's complicated. There's problems with it. But here's what we're doing. And the nice thing about it being community-centric as opposed to just vendor, if this was a feedback to a specific vendor, they would try like crazy to shut that down and get that off the press. In this case, you're gonna see the community start to rally around that and go, how can I help? Where can I write code? Startups will jump in. So really early days for Container, it's got that weird balance between super hyper interest. People are playing with it and it's gotta get mature and there's still a lot of problems to sort of fix. Yeah, absolutely. I mean, we've got the new foundation that was put in place. I think it's OCI now. It was originally OCP, the Open Container Initiative. The ecosystem, I've never seen an ecosystem build so fast. I mean, big companies like Intel and Google and Microsoft throw in a ton of resources. I mean, Satya Nadella has been tweeting about containers and dockers and boy, is Microsoft been moving faster than I've ever seen them move on a technology. You brought up some of the red flags there. Shopify, a company that is presented at DockerCon. We interviewed them at theCUBE at DockerCon. They were one that said, hey, we did it in production and boy, it was tough and boy, we're having a tough time finding peers to actually work with so, but I agree with your analysis on this. We need to see everybody rally and it's not a company. It's that whole initiative. It's really nice to see even the networking was a big push at DockerCon. It's in beta and it's really just, here we're setting the APIs so that all of the other companies, traditional guys like Cisco, startups like Weave, and a bunch of others can really come in and help solve that networking gap, which was one of the two big gaps it was networking and security. The GSA did some things with Docker at the show, along with Booz Allen, but from what I hear, I mean, security still, that's a red flag in my mind and a limitation to what we're offering. You can of course put Docker in a virtualized environment which helps it, but what do you see as the red flags for Docker? So I don't look at it as a specific Docker. I look at it as containers. And I think there's a couple of good things. So you talk about OCI, which is the new sort of standard. It's called RunC, which is the standard itself. It combined what CoreOS had proposed, what Docker was doing. To me, that says, great, we have one consistent standard. I don't have to worry about that fragmenting. And that's good for a couple of things. It's one for people that just want to do containers and it's really important for people that are building platforms. So the PaaS types of companies, the Cloud Foundries, the OpenShift, the Apprendas, Epseras and so forth. That's really good. The other thing I think it does is once you have standardization of the container, then you can think about security. So I think I've been very impressed with what VMware's done with some of their projects, Project Bonneville, some of the other ones that have come out that have sort of said, what do you do if you want to put a container in a VM? Is there something a VM can do to help make it run fast, do security? But I think it's also going to help those platform companies say I can now double down and focus on security. I don't have to worry about the standard fragmentation and stuff. All right, so Brian, you brought up the last topic I want to cover today, which is the platforms. Where are we with the maturity? Where does that stand? We keep throwing away. It used to be we talked about PaaS. Now we're talking kind of cloud native and platforms. Yeah, so PaaS is funny. It goes through these cycles. It was up a few years ago, then it was sort of down the last few months. I mean, we've seen a huge turnout at the Cloud Foundry Summit. DockerCon had a huge turnout, which was the foundation for a lot of these PaaS's. Apprentice just took a huge round of funding. ActiveState just got acquired by, or Staccato, the assets of ActiveState just got acquired by HP. PaaS is back to being sort of hot again. And of course, we had the cloud native, cloud native computing foundation got announced. And that's all about these new applications, which is what PaaS is really there for. So it's back up. I think the nice thing is it's back up in hype, but it's also there in terms of technology. The Cloud Foundry Foundation is doing really well. They're adding more users and more customers. So PaaS is on the upswing. I think the realities of PaaS are there. The red flag for PaaS, not so much a red flag, but just timing, it's a big change. Cloud native applications are a big change. Cloud native organizational models are a big change. So people are going to have to be a little bit patient as to how fast that takes off. And be careful not to get wrapped up in the Docker hype versus the PaaS deployability. Yeah. If I think back to the first conversation we're having here about Flash is, well, how much of the market does that take? Well, Flash is revolutionizing the array with the all Flash arrays. It's infiltrating other parts of the stack. It's affecting cloud. If we look at platforms, how does that fit in the overall cloud market? We're putting out the numbers from Wikibon as to how PaaS fits into the overall market, how it fits into infrastructure as a service. Give us a kind of a thumbnail on that. Right now, just from the numbers, you know, high level, right now we still see it as probably in that 5% range of that overall cloud market SaaS, IaaS and PaaS. But we see it growing. We see it growing to probably closer to 20 or 30% over the next three, four, five years. So it's going to have an uptick that's going to pull along infrastructure as well. And it's got so many strategic elements to it for end customers as well as vendors that it's going to be a fascinating space to watch. All right, can you put together also, if OpenStack is going to be a predominant deployment, at least, you know, growing in the private space and a lot of the platform solutions, like Cloud Foundry are going into private environments, do those tie together nicely? Yeah, I think we're going to see a mix. We're going to see some people that, again, there's rarely ever green fields. So you're going to have some need for virtual machine infrastructure, IaaS. You're going to have some desire to do things, maybe at a lower cost with containers. Yeah, we're starting to see Cloud Foundry supports OpenStack underneath that OpenShift obviously does. Red Hat's made a big commitment to that. IBM's made a big commitment to that. Apprenda is kind of neutral. So I think we'll see a mix. We'll see a mix of OpenStack being the underlying infrastructure, containers being the underlying infrastructure, makes it a little complicated for customers to have to think about all that. And that's why I think we're starting to see more kind of package solutions, because the vendors want to get it out there, customers need it for their applications, but the infrastructure is complicated still. All right, so Brian, if I run through the pieces, you know, one of the areas I'm spending a bunch of time is kind of the converge and hyper-converge space and how that's fitting into the service provider and cloud environments. What's hot on your plate? What's exciting you either in that list or outside of it? Big things for me right now is platforms, sort of this divide between unstructured architectures and structured architectures. We just wrote a piece on that. These new evolving underlying container architectures. And then for me is what's the reality of hybrid cloud? Is it real? Is it really getting deployed? Is it really more public? And people are hoping they can connect them together. So those three areas are going to be top of mind for me. All right, but they're all going to run our watches soon. Absolutely. Internet of things, everything else that goes with it. All right, I think we've covered enough here, you know, a good rundown of what we have. So Brian, thanks for joining us. Thank you. And thanks everybody for watching this rundown of hot technologies and the cold realities. Check out wikibon.com for all the research. Check out siliconangle.tv for the shows and videos that we're going to be doing. And thanks so much for watching.