 Live from Las Vegas, Nevada, it's the Cube, covering EMC World 2015, brought to you by EMC, Brocade, and VCE. I'm focused on Extreme IO, and I'm really happy to have two guests here. First of all, Jane, sorry, Andy Fenslow, who is a senior marketing director. So how's it been, Extreme IO? I saw that you announced some really exciting results and gotten to have you well placed, so tell us about that. It's been magic, and obviously more magic for our customers, but also magic for us. I think it's been really the fastest growing business in EMC's history. We think it's probably the fastest growing business in enterprise storage history, and it's really all about how we are transforming our customers' workloads and how we're really driving broad workload consolidation for the Agile data center. Excellent. We've also got with us, James Moosa, who is manager of storage services, that's by the company in Carolina, so welcome to the Cube. Thanks, David. How's it been with Extreme IO? It's been a great ride for us in terms of not only performance, which is where we began looking at it, but in terms of agility, some of the things that this technology has allowed us to do in terms of freeing up people's time, particularly my staff's time, but then when I look out at the customer base, some of the things we've been able to do to bring healthcare decisions in and much faster have been amazing with this technology. Let's step back a little bit for a second, James. We're a health system located in Eastern North Carolina. We serve 29 counties there, about one and a half million people. We've got nine hospitals and 95-plus clinics scattered across that geographic area. Much of it rules, some of it not. We have a big Level 1 trauma center, which is where our data center is, and this is where we began a partnership with EMC years ago to deliver our electronic health record, which is on effort. And that was the first point that we were driving after to use Extreme IO was to solve a solution for us there to increase performance, give us better space, agility. We wanted that space to be able to grow and take care. And the thing we were looking for was that dedupe and what it was going to do for me in terms of ability to scale. Right. And what were the applications that you're running there? And where were the initial problems that you had? David, my first point was with what I'll call the epic clinical side, where actually the patient data resides. We addressed that first with our first extreme purchase. We immediately saw even more benefits than we could have imagined. And so our second purchase was for the epic reporting side, which is a totally, basically a whole separate infrastructure that you have to put in place. Those people have been on spinning disks since the original days of Epic, so you can imagine those customers when they came live. I mean, I literally had people coming to me going, my report has to be broken, because there's no way it can have completed in that amount of time. You miss some data. Yeah, something happens. No way this can be right. And then you take those people, and so here to four, those guys, here's another thing for that set of people. They had no test system, literally. So when we did an upgrade, we had to take care of a conventional upgrade on this technology. Well, now with Extreme, you've got your old copy. We'd start doing snaps, upgrading those snaps. Imagine how happy they were with this last upgrade. They had three test copies they could play with that only cost me like 30 gigapiece. Those are the kind of things that have just been great for us. Your first implementation. What was the catalyst to do that? That's a great story too. We had budgeted for a VMAX, and I was actually running ahead of schedule in terms of consumption. I was like, I'm not going to make it to when I wanted to get that VMAX. And now that I've looked at Extreme, I'm not sure it's the right play anymore. So I made a very passionate pitch to my CIO and said the right thing for us to do this time, I believe, is to go newer technology, better technology, things that are going to make us or give us the ability to do things we've never thought about. So we brought the first one for the Epic Clinicles. Then we made a major purchase that allowed us to put the Epic Reporting and now other apps on as well. May I ask, when you pitched your CIO for the next set of Extreme I.O. expansion, how was that different from the first pitch? The first pitch was pretty impassioned of, we're going to run out of space if you don't get me something. The second pitch was more, we had run through our testing, and so they had seen, and we've been able to report all the benefits, particularly in terms of things like latency that never climbed regardless of the number of hours. And so it wasn't a very hard sell at that point to say, hey, we need to move totally in this direction, and she's pretty data driven when we show her the data. She's like, absolutely, I agree with you. What does this mean to your people that work for you and your other colleagues in other sections? What does it mean to you in terms of bringing them along, and what does it mean to you in terms of organization? Are you organized the same, or have you had to change that? We're still organized the same, but what it did, we take care of people in healthcare, and for me it was a big deal to see what it could do for people. So if we talk about the people outside of AS, think about these people that were consuming those reports on that reporting side. We literally had a month in process that had two very critical reports that prior to Extreme ran in excess of 41 hours each. 41? 41. 41. They now run in 45 minutes. So think about a clinician now being able to make a decision two days faster, you know. I mean, to me, that's a clinician report. That's not an accounting report. Exactly, right. So it's somebody trying to make a decision about how we staff, how we change our patterns of care, all those kinds of things. They're able to get their hands on that data that much faster. From my guy's perspective, baby, it's been amazing the things that we used to have to staff at night and do, the things, the copies that we had to make that consumed space, we used to have to worry about going to the boss and going, not really sure we can spend this money. I don't worry about that anymore. Why don't you have to worry about it? Well, it only consumes a very, a snap consumes nothing. So I'm not adding space, you know, and I'm not having to do that. And that's a separate copy, fully writable copy. We have, in terms of Epic, David, we have 23 copies, 23 environments. 23. 23 copies. That's a big number. And many Epic customers have a lot of that. So we didn't just put the production copy, we put test copies on there because of what it does for our people process with NIS. This last upgrade, they run you through a transparent upgrade process. All of those copies were done with extreme ids. So my guys that used to have to stay up all kinds of lights and weekends were able to run these things, get them done in just a few hours. And when the app teams were finished, tearing that down and starting over again was a piece of cataconic strength. For many of our customer discussions, we go into the discussion and their mindset is how many copies can we fit and how many copies can we afford. And then our goal is to pivot that discussion back to how many copies do you really want and how frequently do you want them because they're now free. It's really only the changed data that takes up any space. So going to 23 is a big deal and you could go to a lot more. Absolutely. David, here's a great example for you from the guy. I've got a new guy sitting next to me. He's our new director of analytics. He has a dream. Instead of letting the doctors have access to that true reporting database he wants to spin off a copy that can be their sandbox. Very easy for us to do with this dream once he chooses which data element. Well, here before we were concerned about if a doc made a he thinks he's an expert at a sequel quiz that makes a mistake was to do the reporting database. I give him the sandbox. If he blows it away with a snap with the extreme I can have it back for him in those times. Right? It's wild. A lot of people talking about benefits of flash and speed that it deploys. Usually the conversation is really focused on the applications. What you are really talking about is this liberating thing that now you're picking up basically a bunch of man hours that you can deploy on other tasks that here for you didn't have. I have said people have asked me what's your next challenge? What's the next thing? From my perspective I really think our biggest challenge right now is how do we think different? This technology puts you in a place and it conquers the speed and space thing in minutes. I mean you're done with that conversation in no time. Right. Next problem. Next problem please. And so and literally my guys are having to go what process might we be able to make better by using this technology? We've changed it from a that's very expensive and we shouldn't too. Why shouldn't we put everything there? And this is really... So the CIO's behind that? Yes. So the real shift in beliefs that you can go to an all flash data center. What sort of timescale do you think that's going to be? I think we're in good ways out there because we've got some legacy apps and some things we have to make decisions about of what stays on traditional and what goes flash because I don't think we'll put everything legacy on flash for sure. So we've got a few design things to talk about there but we're definitely moving in that direction. In that direction? Yeah. So what about Epic itself? Epic has been around for a long time. What advice would you give to Epic about how they could do things better or differently? Or is it that you can just... The application is fine. Could they actually extend the application? Could they do things that they just couldn't do before? I think they can from a perspective of that application is very dependent upon read-write speeds for success. So much so that on older traditional stores you had to do a lot of definitions of a per spindle basis and how many rate 10, how many rate five. This takes you out of that discussion. You don't have to do it anymore. I've actually spent a lot of time with Epic in Wisconsin and I think the major... You've got the great hair to show. Yeah, exactly. And actually they're an extraordinary company and they've done some extraordinary things for their customers but some of the feedback I bring from their customers. One, I think many customers now see this as the perfect time to re-platform. Many customers may be on legacy Unix or other systems and this is the time to go to virtualization, to go to VMware, no risk, no more concerns about meeting your SLAs. The second piece is a lot of customers really want to get full consolidation. They see the copy services benefit and they love to consolidate production and non-production and that's something that Epic's very conservative about as they should be but they're beginning to think about, is that possible? Now that we have an all-flash array that can really get that consistent performance all the time, no matter how many copies are running. One of the things in health services always is you have to get agreement that from their vendor itself, from the software vendor that you can run it in certain configurations. Was that an easy thing to get? What's it like at the moment? And you were pretty early too. Yeah, we were pretty early. We had a call with them and described what we wanted to do and why and they said, we'll certainly support you. We would like for you to do some extensive testing first. Was this before or after they knew how to spell Xtremite? I'm not sure. They had heard of it. So our partners, Vero and EMC were at the table with us through that testing, Epic provided scripts and said we'd like for you to run through these. My guys ran through them in a very structured manner where literally it was run through this script with these parameters report back to us. Now run through this script with these parameters report back to us. We did that iteration. We went through it twice over the course of about three weeks. Phenomenal results. One of my guys who was a little bit skeptical initially running through the test came to me and actually asked could we put it in production before we had even agreed to because he was so blown away now. The latency never went up regardless of how we tuned the IOPS. And he could see that as a benefit for himself and his team. This is going to be a big deal for us. Right. It's really interesting when you really know it's a significant technology shift when it not only provides you an improvement of performance of what you were doing but really causes you to rethink A, oh my gosh what do we do with this? And now, should we be approaching things completely different? And I'm just curious in your internal life development have you kind of started looking for or identified some opportunities that now we're going to think different based on this performance and capacity that we're really going to approach your problem in a very different way. We certainly have there's a financial modeling app that we moved to Epic on the production side just recently. We found out that this group of people in finance is going to have to start doing a lot of modeling around outpatient care with all the changes in health care. They've got to have a way to model it's got to be different than the way they've worked. So to date one copy per year kind of thing now I'm going to move their test system to extreme so that I can take advantage of the snaps and allow them to do basically the same process that Epic does where they can have updates more often test their modeling processes more often before using the production data. So if this product allowed us to be able to make that change for them. And this is a key example because it's I think most customers are looking at flash as about flashes fast flashes about performance and and that's really just the beginning and with extreme myos very unique architecture and the EMC solution around it it's it's not about performance it's about consolidation it's not about the infrastructure it's about the application and it's not about the workload the individual workload it's about the whole workflow and the business process agility. So last question what advice would you give to somebody in your position what advice would you give to them? If they are in the throes of a storage strategy decision they need to think of it that way they don't need to think of buying flash just for speed and performance think about the potential to bring in all flash moving towards an all flash data center David and making that key and paramount and trying to focus workloads that way great great story Andy Jim thanks for stopping by we're getting the hook we gotta go I love the flash stories because I think you did early some early studies about the adoption of flash and it seems to really be just going significantly faster it's very lonely when you're saying this is what's going to happen and all the all the disc guys are saying over my dead body here we are this lawyer that's all I can say and that will make that our last word I'm Jeff Rick you're watching the Q word EMC world 2015 we'll be right back to the next guest after this short break thanks for watching