 from the Computer Museum in the heart of Silicon Valley, extracting the signal from the noise. It's theCUBE covering OpenStack Silicon Valley 2015. Brought to you by Morantis. Now your hosts, John Furrier and Jeff Rick. Okay, welcome back everyone. We're live here in Silicon Valley. This is theCUBE, our flagship program from SiliconANGLE Media where we go out for the events and extract the signal noise. I'm John Furrier, founder of SiliconANGLE. I'm with Jeff Rick, general manager of our CUBE business. Our next guest is Nettie Shalom, CTO founder at Gigaspaces, welcome to theCUBE. Thank you very much. So we're here in OpenStack in Silicon Valley. You guys have an office in Israel, New York and here in the Valley. So I got to ask you, OpenStack, ready for prime time? Are people using it for mission critical? Infrastructure, what are you guys doing? How do you see all this shaking out? Okay, so that's a trick question right out of the back. So the first answer is hell yeah. It is ready for enterprises. But it comes with a cost and price and challenges and all those other things. Actually, I had a podcast last week around that if OpenStack is ready for enterprises and we had a talk with Lorraine from Forester where we actually analyzed that question and tried to put more color into it. So the reality is that I think the main blocker for OpenStack to be more well adopted within OpenStack is obviously the complexity of the infrastructure itself, a lot of moving parts, a lot of knobs to fix. I think there are a couple of approaches to solve that. Randy Baez actually did a nice interesting comment yesterday is that we all want to avoid the lock-in and therefore we want the options. But we don't understand that but that comes the price. And actually probably the best approach is to use appliances or prepackage type of solutions. And what I'm seeing other companies are doing today like Rackspace and IBM and other private hosted kind of models in which someone outsourced the problem to someone else who has the expertise and deal with that. So I think that if I look at what's happening in OpenStack right now, there is a lot of evolution to create those choices, not just take OpenStack for an open source and build it because most people don't have the skillset, nor do they make sense for them to do that, but actually take it in a prepackage model, outsource model, outsource to the public, outsource to private hosted, all those sort of things. This is going to be in my view critical to the adoption and success of OpenStack in the enterprise right now. Especially since the catalyst of the whole thing was AWS in the first place, right? Dial it up, drop your credit card in and you've got compute power. You think that that would really be a driver to get OpenStack, it's never going to be quite that simple but really be driving hard towards that at direction. I think that's the only solution in my view. Not just installers where I think the primary focus has been, but really create that installed version but appliances, private hosted, as well as public. And actually for smaller companies to consume OpenStack, public would probably be the only option for enterprises that are mid-sized, the private hosted would be the main options and only the big ones that have a lot of engineering forces would actually do the other ones, but those are just few. There's a lot of engineering involved in DevOps, right? DevOps is a great thing, we love it, Agile, you know, you chip in code, we just shipped some new code this morning for one of our DevOps projects and it's a great ethos, we can't deny that. However, most customers have a lot of old stuff. They have servers, they have storage and a lot of stuff's been happening over the past, say six to seven, eight years in memory, flash, converge, hyper and converge infrastructure, seeing VMware out there, a lot of the enterprises. So complexity is there but the price is also going to be associated, as you mentioned, but there are economic advantages. Can you share your thoughts and opinion on some of those dynamics? How do you balance the innovation on the infrastructure side, under the hood? People are building internally and then now later, oh, Kubernetes containers, oh, it's great, love and peace, but in the day, there's realities. What's your take on that? That's an excellent segue to something that I've been pitching quite a bit for a long time, which I'm quoting, the only constant is change and I think it shifts the way of thinking on how we deal with that type of reality. So we need to accept the reality that the only constant is change, that we're going to have new tooling, we're still going to have, especially from an enterprise legacy system that not going to go away that quickly, and we're going to have the next big thing that is coming and actually coming faster than used to be. So from an enterprise perspective, the problem becomes even bigger because this environment is going to be more and more heterogeneous, you would not be able to see an enterprise saying today I'm a VMware shop or I'm an Oracle shop or I'm something shop because it's actually going to have everything in its tool chain if he wants it or not because that's the thing for innovation. That in my view, shift the value as well as the problem to more of an integration problem than it used to be. So we cannot expect from a vendor to come with the monolithic solutions like we use with Tivoli and IBM and other things. Okay, let's, yeah, I'm your guy. He's going to figure out how to do all those things for us and we'll buy it from someone else and he's going to give us a pre-packaged solution for everything and now it works. That's not going to work because obviously that vendor would one lock you in, second he's not going to be able to embrace that change himself. So I think that right now the main challenge is how do you integrate all that? How do you create an integration tool with existing stuff and new stuff? Exactly, that you could yourself without being dependent on a vendor or plugging in those new stuff but it would look from an end user perspective from the users that need to consume those stuff as a holistic solution as if it was pre-integrated and that's something that for example we're focusing on. You mentioned appliances earlier in your early comment. Okay, that was, people were against the appliance. Oh, I don't want to know the box. That was kind of like, you know, sprawl that conversation around server sprawl going back, you know, a decade ago. Consolidation, virtualization, certainly changed that game but now you're seeing the trend tip back to as Oracle says, engineer system. So actually not a bad thing. Like I plugged something in, it works. My iPhone works. Android kind of flakes out a little bit here but you know, that's kind of the mode. This iPhone kind of mentality where hey, I want to drop something in. I want a bulletproof. I want an SLA. I want the scale. I don't really care to the box. That's what customers we hear from customers. Do you agree with that? How do they manage that? The reason why people were scared away from appliances is because of the locking concern. But if the appliance itself takes an open technology and prepackage it for me and do the same thing that I could do myself, then there is no harm doing it because what value would you get to do it yourself? Actually, no value. And I think that's the balance. And I think as long as we get appliances that relies on open technologies and do the same thing that we could do it ourselves, as long as we keep the net line and there's nothing wrong in appliances, and that mindset needs to change, that negative mindset towards the client needs to change. I want to get your thoughts on this because it's a great conversation because this is kind of like a, no one really talks about it in public because it's hard to articulate, but back in the old days of software, they had the Intel processor that was a hardened top, a lot of proprietary stuff and a processor, but no one really cared because it worked. So back to the kind of engineer systems and this middleware battle going on, call a pass, call a middleware abstraction layer, containers, DevOps, that's a big focus. It's new middleware markets evolving. So the question for you is, where's the hardened top? Where's the line where it doesn't matter, it works. Is it the infrastructure, is it where in the stack or can you add some commentary to that or insight? So obviously I have an opinion about that and I'm biased towards the approach that we've taken, but in generally, I think that the underlying infrastructure itself and tooling is going to keep on changing very fast. The thing that changed less is the application that we consume and use and the thing that changes how do we run it, where do we run it, how effective it is and we also want to be able to be able to push into those applications the new features that we want to do but that's a layer that we usually control and that's something that we can embed and embrace. So the layer of complexity are the things that we don't control and usually that's in the infrastructure there and therefore I think that the right level of obstruction is really closer to the application rather than in the lower level of the infrastructure because that's going to keep on changing and moving and be disrupted quite significantly and we want to keep it as close to the application which is a layer that we can control. So if we put the obstruction at that layer, our ability to actually create if you'd like a shock observer between all the things that are moving and the things that we control, then the chances that it will be sustainable and stable over time is much higher. Then if we try to put the obstruction at the infrastructure level, at the storage level, at the network level, all those type of things. So if you believe that to be true, which I like that idea, let's just play with that idea, the opportunity is pretty large for people, especially in the software business. What kinds of opportunities are out there for other entrepreneurs, other companies to innovate because if you assume there's a lot of change going on in the underlying infrastructure and we're seeing things with Flash and whatnot. Yeah, so I mean, I'm inspired from the movie here, have you seen the movie here? H-E-R? Yeah, yeah, oh yeah. And X Machina, I'm not sure if you've seen X Machina. Yeah, so if you look at that, I think that's where the next wave of things gonna push forward, which is combining VI, intelligent and all those type of things into infrastructure and orchestration, which in my view is the body and the brain. So right now the body is built separately and the brain is built separately as we as human are mediator between the two things and plug them together. So we need to see what the infrastructure speed out and then say, okay, now you do this, move your hands here, move your hands here, which is kind of broken if you think about it. So the main point of innovation is how to glue those two things together to create this what I would call a software defined operator, which basically means robots. It's very similar to what we're seeing right now with self driven cars, autonomous cars if you'd like, where we used to as humans either hold a driving thing, but we know that as drivers, we're not that good and we make a lot of accidents and those are areas in which you could innovate quite a bit. And I think that's the main area of innovation. How do we take a lot of that manual work, a lot of those things that are fairly complex, get put the robots inside and get them to be. I think that's great insight. I mean, I think if you look at how people would take an approach to marketing, oh, here's my silo. I'd build this software stack and it's a silo. But what cloud has really shown us is it's horizontally scalable, right? So like, that's a beautiful thing. That could be the head, right? So if I've got an app on a self-driving car, you can have Intel processors and all that stuff in there. The software's going to run on a self-driving car, my watch, data center. So this interoperability concept is not just about interoperability, it's functionality too. So it really is vertically oriented stacks with a horizontally scalable. Do you, is that kind of, might get in that right? Yeah, yeah, because it does drive the focus. I mean, we're all talking about speed, right? That the focus needs to be on speed of how fast we can roll out new features, new applications to be more competitive. Again, similar to what Ford did when he roll out Model T. If you think about what he did back then, we used to build custom cars before that and we were very proud about that. And then Ford came in and he said, you could choose any color as long as it's black. And shifted all the thinking, when you scale, you have to think of how fast you could roll out things, not how special they are. And we go into the same point with IT industry. We were building data centers by hand and now we're changing the focus on how fast we can roll out things. And over the sudden, being special is actually a barrier. So if people have, if there's going to be no winners, basically a coexistent model means there's no winners. I mean, there'll be a bigger winner than other winners, but if you have a lot of things going on in an enterprise, we vertically integrate horizontally scalable, that speaks to the trend of purpose built stacks at Oracle, you see VMware and EMC, that whole conversation, that could go end to end. So this notion of end to end data center to the car, to the edge has to be not monolithic, one vendor has to be multiple vendors. By destination. So how does OpenStack win there? Cause that's really, I'm trying to see that. I just, I'm not seeing how OpenStack can be an end to end winner with all the agendas. Yeah, so I think that's a gap that I think exists right now in OpenStack. If you, that's the paradox I was writing about the OpenStack paradox, interoperability paradox. That's exactly that thing that I saw as well. You published that? Yeah, I just published it yesterday, I think. Okay, on your blog company blog? On your company blog, which touches exactly on that. When people buying to OpenStack, it's for that interoperability, but the reality is that they cannot use OpenStack for interoperability because it was really built to be a clone of Amazon on a private cloud. And there is almost zero focus on interoperability. That is changing, but I think right now, it's the ecosystem who solved that problem. I think there is a lot of solutions now around OpenStack to solve that problem, including what we're doing. But I think that focus on OpenStack should change from how do we build the next Amazon to how we build OpenStack so that it could run on VMware. So it could run, which actually does that already, but could also run on Azure and Google and those type of areas. I think that's the area in which OpenStack would thrive and be much more successful than if we take OpenStack and build the best storage and the best network or the best whatever. So if I think about the two scenarios, OpenStack now runs as the infrastructure or at least as abstraction across all those clouds versus it builds the greatest computer storage, which one are better? I think it's there. So OpenStack is at a inflection point then right now. I believe so, yes. And it's like a moving train. It's like shooting a moving train, right? It is. And the $100 million of Neurantis, I think was really targeted to solve some of that change and at least bet on one company to strive that because I think you need the powerhouse of developers to focus only on that, to actually drive that because the rest of the ecosystem I think is less if you like focused on solving that problem. So I think that's a good news in this event and I believe we'll see a change as a result of that. Final question for you. Is there a hybrid cloud? Does it exist? Is it a category? I mean, you got public cloud, you got Amazon, you mentioned that. Private cloud, we're seeing that. Those are tangible solutions. People are deploying. I mean, hybrid cloud, does it exist? Is it like, do I buy hybrid cloud? Is it just an outcome of deployment? So, the answer is yes. It exists by definition even if you look at organization you'll see that some things that run on Amazon but some things they still don't run on Amazon but it's not a solution, an integrated solution. It's a reality but no one figured out how to make that reality a cohesive experience, integrated product or whatever. And again, if you look at any enterprise you'll find them using not just Amazon they will be using something else as well and they will be using VMware as well and they will still be using VMware. So how did it exist? Just not there as a solution, not as an architecture, not as a product. I think that's kind of the gap that we're talking about. It's kind of like distributed computing, it's out there. Exactly, it's by definition. I mean, you're using iPhone but in your organization they would be using Android as well and some other things. So by definition it's going to be even more and more hybrid just what we haven't figured out how to make it a cohesive architecture that hybrid approach that becomes around. Thanks for your insight. Really appreciate you taking the time. Coming on Natty, the CTO of Gigaspaces founder doing business mission critical applications in the cloud. Obviously the heart and top is close to the applications. Everyone's going to be winning to what level we'll be following that here in the queue. We'll be right back live in Silicon Valley after this short break.