 Now, here is your host, Dave Vellante. We're back at Oracle Open World here at Moscone Center. We're in Moscone South in the QLogic booth. Please stop by and see us. This is theCUBE. theCUBE is SiliconANGLE's live mobile studio. We got to the events. We extract the signal from the noise. This is our fifth year at Oracle Open World and we were able to get here in 2010. Really, there were two companies that enabled us to be here. First is QLogic. QLogic gives us this great space within their booth, very generous. The second was EMC. EMC at the time said, we think that the Oracle customer base is important. It's one of our biggest customer bases. It's strategically important for us and we want to help our customers understand what's going on at Open World. So EMC helped us get here. So thank you for that EMC and QLogic. Vinay Gankar is here. He's the director of product management for Extreme I.O., which is a division within EMC. So we're going to talk about Flash. Flash is one of the biggest disruptive changes in infrastructure. It's got a huge affinity to databases generally and Oracle Database, specifically Vinay. Welcome to theCUBE. Thanks for coming on. Thanks a lot Dave. Really happy to be here. So it's been an interesting week already. You know, Sunday night, Oracle comes out. They got the slide. We're X times faster than Extreme I.O. We got this new product. And it's quite interesting to me because it's sort of an apples to oranges comparison. Extreme I.O., and I believe it's called FS1, is the Oracle product that's a sand that has spinning disk in it, it has Flash in it. Do you have spinning disk in your product? I don't think so. So what do you make of all that? Hey, great question to start off. It came out last night. We don't have all the time to look at it, but reading through their marketing materials and datasheet, it looks like a hybrid array to me so far. One of the key things that kind of came out after combing through their datasheet, there's no mention of latency in there. It's really interesting when you talk about Flash, Flash starts with latency, right? So to me, it's a hybrid array with Flash and disks. So we have to see more data in terms of, you know, what exactly it can do and stuff like that. Yeah, I mean, you're right. It's surprising that Oracle's not talking about latency, especially given that we're talking about a database environment. Why does latency matter in a database environment? Can you explain that to our listeners? Yeah, absolutely. Think about Oracle. You're running your OLTP environment. You're running your customer entry or customer application on your database infrastructure. So every microsecond that you can save has huge impact on your users or your customers or your business processes. Latency directly transformed into the business value. The amount of money you can save and the faster process you can complete, right? So latency is an important aspect for certain Oracle applications like OLTP kind of databases. So that's why you see that Flash has been this transformation technology, especially for databases. So to me, the number one characteristic of Flash array has to be this latency, which is consistent and which can be consistent when you start loading your system, when your data growth increases what you have when you start loading up your system. So latency's all, Flash is all about latency to start with and there are obviously other things that Flash has to do. Yeah, so that's how I get, DBAs want an application as they want consistent performance. That's what they always tell us. You know, raw performance, that's nice. The benchmarks are nice and I love benchmarks because it's a stake in the ground but it's that consistent performance in a real world application. So my question to you is, give us the update on Extreme I.O. What are you seeing in terms of your ability to affect that consistent performance, that desire for consistent performance? Sure, absolutely. So one thing we see trend that's happening is in the database environment, now the DBAs take the center stage. DBAs wants to have all the productivity tools in terms of reducing their complexity in terms of managing the database infrastructure. Like a recent I.O.U.G, independent Oracle user group said, DBAs spend about 70% of their time looking at performance issues. They're not able to spend enough time on strategic issues like, can we consolidate? Can we bring more efficiency to your infrastructure? Unfortunately, they're not able to do it. Now with consistent performance, you're taking away the performance issues associated with your database, right? That's number one. So as part of that, you have to make sure that you can deliver the consistent performance while delivering other storage capabilities that you have. For example, deduplication. On Extreme I.O., deduplication comes as part of the architecture. Running deduplication does not have any impact on performance. The consistent latency remains consistent. We also have Snapshot, which we released about a quarter ago. Snapshot have almost zero impact on our storage performance. And then, and we're going to announce, we have announced as part of an upcoming 3.0 release of compression, right? Now together with consistent performance and data reduction technologies and data copying technology like Snapshot, we have a complete solution for databases. And if we talk about this to any DBA, the DBAs like this story very well. We have done so far very well in the few quarters we've been selling. So it's a great ride so far. We think we have a very, very strong, effective, interesting product in Extreme I.O. for databases, Oracle databases. I've seen that Oracle user group study. The independent I.O.U.G. Independent Oracle user group study talks about how DBAs spend their time. We actually went out and talked to them and some of the ones that had brought in a converged infrastructure to understand how that affected their time. Whether it was an Exadata or a Vblock or whatever, who's ever converged infrastructure and studied that. And we found that it dramatically impacted their time positively. And they spent now more time on the application. Have you seen, do you have examples or proof points where folks have brought in Extreme I.O. And it's had a positive impact on the time they spend doing non-differentiated heavy lifting? Absolutely. I can talk about a big medical analytics company in the New York area. So these guys manage one of the biggest medical analytics shop in the United States. In their case, they have production databases. The production database, in that case, is kind of OLTP and also data warehouse. It kind of merges your traditional OLTP and data warehouse, which means that you have small, random I.O. and large, you know, sequential reads, right? So that's one aspect of the production database. So they moved that to on Extreme I.O. And they're seeing huge benefits. Now that is kind of expected that we can provide the consistent, you know, low latency, consistent performance. But in addition to that, they have a huge ancillary infrastructure, which is they're testing there. There are other infrastructure that are needed to be making sure their production runs fine. Their new applications are tested, right? So now they're able to bring all that into one infrastructure where they can manage, they can consolidate the whole thing, they can manage it more in a simple way and also brings out more agility because their infrastructure is not one infrastructure. They can create copies on demand. They can, you know, finish their testing and development more quickly because they're testing them, now runs on their production system, which is same as production, unlike in previous cases where they had to isolate this to make sure their production runs okay. But because of the consistent performance Extreme I.O. provides, they don't have, that is not an issue anymore. So it's like, and then for this, there are other customers like this, we're seeing on a day-to-day basis now Dave, so it's amazing to see this story is coming in. As you probably know, early on Wikibon did some research on how to sort of optimize your Oracle license spend and what we discovered was, and this was sort of pre, before the days of all flash arrays really exploded, but we started with taking Flash, putting it into an existing hybrid array and optimizing the I.O. infrastructure. And what we found was, you could dramatically cut, I mean, I think it was up to 30% plus in the cases that we studied, the number of cores, which had a direct impact on database licenses because Oracle licenses its database based on the number of cores, which then ripples through to the maintenance. So I wonder if you could talk about that, what you've seen in the customer base. Of course, not just Oracle, it's true for SQL Server, it's true for IBM DV2, which just makes sense, that's how they all price. So the bottom line is beefing up your storage infrastructure a little bit, maybe spending a couple hundred thousand dollars more actually could save you significantly more than that. 500, 600, 700, maybe a million over the life of those database licenses. Have you seen that in the field? Absolutely, we see that almost every day. Now, the reason for that is that the typical sizing for cores is based on what is your storage can provide, the latency it can provide, right? So- Oh, we're talking about latency again. Exactly, I mean, it finally comes down to latency. Now, when you talk about that, the latency impacts the utilization of the CPUs, the cores, right? So now, if you can provide better latency, you can make your cores run in better utilization than it used to before. So, which also translates to reduced number of cores. That's the, you know, that's a natural way of thinking about. And that's how our customers are thinking, can I run my 100 queries that need, let's say 20 cores today or 24 cores today in less number of cores? That's the question they're asking. Now, because I have faster storage, I should be able to do that. And that's been the proof themselves. So obviously, like you said, it is reduced number of cores, right? It's kind of very obvious how that happens. And also all the other costs associated with the core, maintenance and what not, right? Yeah, this is a very interesting solution flash, especially for XTMIO in terms of how you can consolidate and reduce overall infrastructure cost. You know, I saw the Gartner Magic Quadrant come out. Obviously, you guys must have been pretty happy with that. I mean, XTMIO is relatively new and you got the top spot. I have to say, I was a little surprised. Can you talk about that a little bit? I mean, I guess, I mean, Pure's been out for a while. Maybe a different class of product. Why do you think you were able to achieve that? I mean, that's a horrible question, but I want to unpack that a little bit. You know, people used to say, well, it's because you spent a lot of money with Gartner. Okay, I know the Gartner guys. And while they get criticized a lot for that, it's not that easy. You know, you got to have some chops. So, but I was a little bit surprised given the newness of XTMIO. You guys were in directed beta for so long. What am I missing? So, we are really happy to be figuring the top right corner as part of the Magic Quadrant. So, Dave, it all starts with the EMC brand value. EMC has been selling storage for years and we've been the leaders, right? XTMIO is the new form of array, new all-flash array as part of EMC bigger portfolio. So, the brand of EMC automatically is now attached to XTMIO. The sense of quality EMC brings to the table and our ability to execute, right? All these different aspects of EMC comes into play when we talk about XTMIO. That's what we're not surprised that we are there because we are EMC and we have a great storage array, all-flash storage array, which is very unique in the industry. Like I said, we've been shipping only for now three quarters. We're not done, this is the third quarter we're shipping. So, we're happy to kind of getting that. We hope to improve upon that in coming quarters, coming years. Now, what should customers know about XTMIO? We've got new software release coming out. Can you talk to practitioners about the migration plan from where they are today for the existing customers? Obviously, the new guys have to have to think about it, but for the existing ones, how do they go from A to B and how do you minimize that disruption? Can you talk about that migration plan? So, I believe you're talking about our upcoming 3.0 release. Yes. So, let me highlight some of the key aspects why we are doing 3.0. So, 3.0 is essentially very relevant to the audience here. Now, 3.0 will improve article performance by 30% to 40%. And it'll cut down storage costs by half because of our compression. That is the biggest value 3.0 is bringing. So, we are offering an EMC service. It's a free service to our customers, existing customers, to move from the existing version to 3.0. Now, our customers have been appraised from beginning. They see the huge value 3.0 brings in terms of performance, in terms of compression that we bring into the table. So, for us, it's about providing them the part and we are doing it through a free service, EMC service. It is like, Dave, it's like you have a Mercedes car. EMC will bring with another Mercedes car and a chauffeur and they will take you for a weekend ride while your existing car is being serviced, upgraded. When you come back, your car is now more performance and more efficient. So, that's the way we're thinking about it. It's a service that we are offering. So, we are really excited that this is coming along and we're really great to see all the input and appreciation from our customers. Well, that's classic EMC, right? I mean, migration's never been an easy thing. So, okay, so you're wrapping services around that to make it as seamless as possible for customers and then these customers get compression, D-juice already there. What else do they get? So, 3.0, like I said, it improves Oracle performance by 30 to 40%, right? So, that's the biggest thing that you're going to get in addition to compression, right? And we think that this is a new platform for us to deliver new services on top of this architecture. So, now your competitors would say, oh, it's a disruptive migration. Can you promise me that this will be the last disruptive migration that I'll have to face or no? Yeah, absolutely. That's what we are thinking about it. So, this was a tough decision there. This was a tough decision to have this what we call upgrade which needs EMC service, right? It was a tough decision, but we had to do it. I think we're doing the right thing because we wanted to deliver more value to our customers. Well, it's early on too. Had you done this four years in, would have been a different story. No, absolutely. And then note that this allows you to keep your existing investment. We're not asking to throw away your hardware that you've already bought from us. So, it'll keep your existing, you know, intact. It's a software upgrade. I don't have to buy new hardware. And EMC has a lot of experience with this, but you know, especially with Symetrics and VMAX, and that was big. So, you learned a lot from that. And you've been able to keep your customers to your point about EMC's execution. People are always telling me, because they always get accused of being an EMC fanboy, but you have to give the company a lot of credit for its ability to migrate through some of these really difficult hurdles. Right. And customers keep coming back for more. Oracle, same way. You know, you got to give companies that are able to migrate to those transitions a lot of credit. And the reason is because the delivering value. So, where do you see this whole flash thing going? There's a crowd out there that says, oh, it's going to replace spinning disk. Greg Shearer was on this morning and said, even if you want to replace every spinning disk out there, there's not a flash capacity in the planet to do that. So, obviously that's not going to happen anytime soon. Having said that, there's really alluring value in flash. All active data will be on flash. So, where do you see this whole thing going in the next three to five years? So, if you look at the data growth day, the most of the data growth is happening in the active data, like you mentioned, right? I believe that my personal opinion is that most of those active data will be on flash. So, which means that there is a huge growth in active, there will be huge flash adoption because to serve this active data. What happens to other kind of data? Obviously, you know, data is made of different kind of different storage out there. People said tape will go away, state still exists in enterprise boundaries. Same thing happened to a disk. But, the amount of huge growth that you will see in active data will really propel this flash adoption. Vinay, thanks very much for coming to theCUBE and helping us unpack some of the innovations that are going on in Extreme I.O. We're very interested in seeing EMC's formal response to the Oracle's claims. That'd be great to see. It's a lot of fun here at Oracle Open World. It always is. So, again, thanks for coming to theCUBE. Thanks a lot. Really happy to be here. All right, keep it right there, buddy. We'll be back with our next guest. This is theCUBE. We're live from Oracle Open World 2014 and we'll be right back.