 Live from Las Vegas, Nevada, it's the Cube. Covering EMC World 2015. Brought to you by EMC, Brocade and VCE. Welcome back to Las Vegas, everybody. This is Dave Vellante. I'm here with David Floyer. We're at EMC World 2015. Live, two Cubes here in Las Vegas. Boaz Palgi is here. He's the Vice President General Manager of Scale I.O. at EMC. Boaz, great to see you again. Welcome to the Cube. Thanks for coming on. Thank you. Well, happy to be here again. The big day for today. We had this keynote day was Emerging Technologies Day Three. And a lot of talk about Scale I.O. It's our great demo from Chad. You must be really super excited. Absolutely. Congratulations. Give us the update. So what we, the two main things we have been announcing this week are the VX REC by VCE enabling customers to consume Scale I.O. not just as a software product, but also as an integrated, converged infrastructure appliance from factory. And the other big news, of course, is the availability of Scale I.O. software in a completely free and frictionless manner for testing purposes, for non-production test purposes. Yeah, that was announced this morning. Yes. If anybody can download now the software. So maybe we can back up and help our, I mean, we've talked about this before in the Cube, but help our audience understand sort of the concept behind Scale I.O. So, if you look at the RIS, there's no data of the same technology today. You need to go back 25 years and look at what the resources were available at the application server at that time. So servers at that time had just enough resources to run a single application. So if your server would go down or your internal disk holding your data would go down, you would lose everything. So people wanted to create some application that would take care of resilience and high availability of the data, but they couldn't run it on the same server because the single application on the server had to call the resources. So what they had to do was take a separate server, a new server, put some additional disks in there and run a single application there that managed the data. So that server, that storage server actually grew into what happens to be a $50 billion sand industry today. It's a big, big, big industry, but not too bad, exactly, but it's a workaround. And so if you look at any server rolling off the production line today, even the local servers have ample resources in terms of CPU, memory, local capacity and magnetic and flash network. So you can today run a product like Scale.io, which is an enterprise class storage system as just another software alongside your applications, databases, and hypervibes on your servers, and thus eliminate the pain really of running multiple layers that you had to do as a workaround in the past. Okay, so that sounds good. What's the challenge? What's the trade-off of that? You've got to be able to recover, you have to be able to support enterprise class availability. How do you solve that problem? So Dave, as you know, our team has been working together in this industry for 15 years and we have been building storage software products a few times, both in startup companies and in incumbent vendors. And so the unbelievable unique of Scale.io is that not only don't you really have a trade-off here, but actually because of the architecture that is a scale-out architecture that parallelizes your IOs, you not only aggregate the capacity across all the nodes, but you also linearly scale performance. So Scale.io really enables you today to get the highest performance storage system in the market today. And actually, if you saw the VX-REC announcement by VCE earlier this week, they showed a thousand nodes environment where you get 38 petabytes, but also 240 million IOs out of a scale or a system. So apart from that, you get all the basic enterprise functionality, advanced functionality that you can expect from enterprise class systems, snapshots, thin provisioning, all kinds of advanced, other advanced functionality, management, APIs, to other management systems, et cetera. Dave, I know you've been a big fan of this architecture. Yes, I have, yes. I've written quite a lot about what I call ServerSan, it's now called hyper-converged architectures, and when we first created our company together, we were doing forecasting on the SAN itself, and why it was good, why it helped to release stranded assets, the very good business case that SAN's had in those days, and they grew very, very fast indeed. And it strikes me that ServerSan, putting it together, actually makes that whole system so much simpler because you're combining everything together. So our forecasts are actually very, very positive on ServerSan because it's a more efficient lower-cost solution. Is that, are you seeing that you are going to be the savior, if you like, of the old SAN business? Well, I don't want to call it the savior, but I definitely think that when you look at this space and you look at the challenges that customers have to deploy a product like this, I definitely believe ScaleIO is the front runner by far with a very long gap behind us to the rest of the field, and the reasons are three or four fold. So first fold, we talked about Scale. So modern data centers today, after all the consolidation that we have seen over the last few years of data centers in the enterprises or in the SMB into hosting and service providers are running hundreds of thousands, if not millions of servers, which is very different from the traditional data centers that we all grew up on with a few tents or a few hundreds of servers. So the capability of being able to run at scale, a true scale is paramount, and obviously ScaleIO, the name says it, the core capability that we put as one of the design capabilities of ScaleIO is we are able to scale the product from three nodes to many, many thousands of nodes. The second thing relates to the elasticity. So when you run these kinds of environments, you cannot tell the customer what to do with this hardware. The customer needs to be able to add and to move and to remove servers whenever he feels like it in a completely liquid manner. He also should be able to have any configuration. So he cannot be forced into symmetric nodes. He must be able to have nodes with different types of CPU and other configurations, number of days, type of days, speed of days, et cetera, and ScaleIO enables that as well. Then the whole issue that they brought up of enterprise advanced functionality, you need to be able to provide the customer, all the enterprise functionalities that the customer expects from enterprise class system and in a very robust manner. So one of the feedbacks that we are getting from our customers is that they are not able to break this product. And now that we put the product out there free for download, it's going to be interesting to see if people are able to break it, obviously within the limits of using the hardware in the configuration that works. Now I presume you're seeing demand from service provider space. Maybe talk about where demand's coming from. I mean Verizon was up on stage today. Yes. Talk about that a little bit. So we see two main use cases in the market. One is virtual server infrastructure. So this is indeed server providers with their public clouds and infrastructure as a service, but also enterprises with private clouds. And we see transactional databases. So SAP HANA, Oracle, Oracle Rack, SQL, et cetera, MongoDB. And so we see those type of use cases taking off because customers in those environments really are looking to lower their TCO, but also improve their performance and their operational flexibility and elasticity. Now the other important thing is that it's easy to say, hey, let's move our technology overnight to a hyper-coverged. When you look however in these large organizations, it's easy to make a technology swap, but it's not that easy to make an organizational swap overnight. And here is another unique capability of Scalao's architecture, is that we enable the customer to actually make this transformation at his own pace. So he can start in what we call a two-layer architecture, where one layer of servers runs the applications and the compute, and the other layer of the same type of commodity application servers runs the storage with Scalao. At this time, you don't need to make any organizational changes. And then over time, when the application guys need more compute, they can just run the applications or the hyper-visors or the databases on those servers that are originally just running the storage, turning that layer into a hyper-converged layer. And similarly, when the storage guys need more capacity, they can just put Scalao data server software on the application layers and turn that layer into a hyper-converged layer running compute and storage as well. And so that transformation from three-layer traditional data centers to two layers, compute and storage, to ultimately a hyper-converged infrastructure without ever having to do any data migrations, without ever having to interrupt user's I operations, is what Scalao again is truly unique. So one of the things we're also seeing obviously at EMC is the great success of Xtreme I.O. And I've been talking to a whole number of customers who are planning now to go to flash only data centers. Something that would never have been considered just literally a year ago and working to it very, very hard indeed. How do you see your product in that area? They've had to put very specialized things into Xtreme I.O. to make it effective as a Scalao architecture, et cetera. How are you planning to go into that world and how are you offering different opportunities for customers in that new world of flash only? So first of all, we are agnostic to the underlying media. So Scalao today works with a flash where there's SSDs or PCIe cards of the customer's choice as well as with magnetic disks, SATA, SAS, et cetera. Obviously Xtreme I.O. and Scalao have two very different design prerogatives. So Xtreme I.O. is a capacity optimized solution with a fan granularity I.O. layout and completely optimized for the best capacity. And obviously applications that are well dedupable like VDI, et cetera, have the best results in the Xtreme I.O. environment. Scalao has a different layout, which is a mid granularity layout and we are optimized for performance. So capacity optimization performance are almost each other's opposites because in order to get capacity optimization you need to do all kinds of redirections, et cetera, which inherently hit performance. And if you have a performance optimized layout like Scalao you don't get capacity optimization. You do get your cost savings by your ability to buy the media of your choice. You can go and buy SSDs or flash cards, PCIe flash cards from whatever vendor you want and you can play them off against each other and get the best price quality flash that you want in a Scalao environment. Angelim is swinging to a software-defined world. Boas, thanks very much for coming on theCUBE. We really appreciate the update, the insights, congratulations on all the progress and good luck in the future. Yeah, thank you for having me and it's been a pleasure as always. Pleasure as always. Keep right there, everybody, we'll be back with our next guest right after this. This is theCUBE, we're live from Las Vegas, EMC World 2015, we'll be right back.