 SiliconANGLE.com's continuous coverage of HP Discover 2011. SiliconANGLE.com is the site for reference point for tech innovation. We have news every day. SiliconANGLE.tv is where we take video out to the audiences and stood theCUBE as our flagship telecast. We go out to events and we explore knowledge and share with you and let ideas grow. And we're on our summer tour 2011 and we're here with Cornell Medical. Welcome to theCUBE. Thank you again. Okay, Eric, Dave, what do you think about the customers that are here? I mean, you've been talking to a few customers. At VMworld, we did a lot of customer interviews. Before we get into some of it, what's your assessment? We talk about proof points. Well, my quick assessment is HP customers have always been very loyal. HP has always scored high on things like customer satisfaction. Now, of course, Eric, you're a three-part customer. Now, you may be an HP customer as well. We are HP as well. So we have a lot of C class blade chassis, blades and blade chassis. And we were originally a three-part customer we've since moved over with the acquisition of HP and three-part. So tell us a little bit more about your role, your organization, what the IT organization looks like. Can we set it up? Sure. We have a lot of different responsibilities as I'm sure you hear a lot these days and we're trying to do more with less. And what we deal with is we have a lot of different people are throwing a lot of different requirements at us. Things like our clinical systems, exchange, virtualization. And one of the things that we were having problems originally with our storage is we were an EMC customer and their model at the time, and I'm sure it's still the same, is we were building purpose-built arrays. So we had one for email, we had an array for clinical systems and then moved that on and on. And we built out six or seven arrays and we realized that wasn't a sustainable model. Why? Because the infrastructure wasn't fungible across the application portfolio? We didn't have, we're in Manhattan. We have limited space. We have limited power. And we have limited human resources in order to manage all those things. So we're trying to get the most out of what we have with the limited resources and the ease of management for things like the C-class blades and for 3-par has been a huge benefit for us because you don't need that many people to actually manage what's going on anymore. So what was the symptom of that? Just there was people, it was... Gapx, Opx, Cooling. It was power, space, cooling and the management aspects of it. So we can consolidate all that down which is what we did into two arrays. We took those six arrays and we consolidated them down into two. And now we basically have two people who manage the fiber channel infrastructure and also the center arrays. So you're supporting all those applications that you mentioned, you mentioned Exchange, Clinical. Are you doing Exchange 2010 yet? We're in the middle of doing a migration so we're on... You're putting a sand on Exchange, come on. We just have Microsoft on, yeah, I'm supposed to do that. No, it just depends on who you're talking to within Microsoft. It depends on who you talk to. I've heard them say that you can do Direct Attach. I don't see a lot of people doing that. You know, I coined a blog a while back that said Microsoft has its head up its dash. Pretty funny. And they were just on the cube but they leave the area? Well, no, so the inside baseball there, John, is that Microsoft's Exchange group doesn't want you using sand on Exchange. They want to simplify. But in reality, we know if you can share that infrastructure across multiple applications, especially for a mid to larger size organization that's much less expensive, right? How many seats do you have on Exchange? We actually have about 30,000. Yeah, so we found that the crossover point was about 750 seats. And I think that shifted a little bit. It may have gone up with Exchange 2010. It might be up to 1,500, 2,000, but 30,000, no brainer from a cost standpoint, isn't it? It's absolutely no brainer. So you sand, I mean. Yeah, and the large requirements that we have for our mailbox users, I mean, we have people who have very large mailboxes, 20, 25 gig, 30 gig mailboxes. And to put that on Direct Attach is just not doable for us, you know. We have a lot of people who would complain pretty quickly if we started using that. So one of the things with three parts is you're able to leverage the virtualization of all the disks. So to spread the IO out across the pool of disks that we have, and that's just not something you can do with DAZ. So that's one of the big benefits for us. And not only do we have Exchange running on there, Exchange is virtualized and running on the same disks that we have our mailbox databases running on. When did you bring in three-par? Four, right? Three and a half years, we've had three-par, so. So, Eric, I have a question for you and I don't know if you can answer this because I talk to a lot of IT practitioners and they know their space really well and they know their products really well. They tend not to know other vendors necessarily really well because they don't have hands-on experience. But, but you saw sort of traditional storage. Now, you've seen that morph to add what we're called by some people, feature capabilities like thin provisioning. Right. To their products, does that tick the box? Does it do the equivalent or is three-par really that different is what I'm trying to get to in your estimation? Absolutely, three-par is different. I mean, in the mid-range, which is where we are, I mean, we're not an organization that's going to be able to make a huge investment in something like a DMX from EMC. The space requirements, the power requirements, just the monetary requirements were just too high for us. So, we graduated a little bit from where we were with EMC's CX500s and we wanted something, we wanted to reach for something a little bit more and we couldn't go for a DMX. So, three-par fit that model for us very well and thin provisioning was one of the huge, huge benefits that we saw right away because we're presenting out terabytes and terabytes out to our host and realizing that they're not actually using that data. Somebody comes to you and says, I want to buy 10 terabytes of storage but in reality, they're probably only using one terabyte or maybe even less than that and they just want to have it just because they want to feel good and they bought it and they want to make sure that they have it. So, what we did with thin provisioning is we provisioned out all that space to them and it looks like they have that space but they don't need it. So, we can use that for other things and sort of build that around our budgeting cycle. What's your organization look like, your IT organization and how does it report up into the broader organization? So, we have different groups who are basically spread across the different technologies that we have in database, email, virtualization and so then we have the CIO up at the top and he'll come down and talk to you anytime the day or he'll send you emails anytime at the night and he'll just come down and talk to somebody like me who is at sort of the mid-level. Just a regular guy. Just a regular guy doing my thing. Who's the CIO report to? He reports to the dean of the college. So, is the dean kind of like the CEO of sorts, right? So, a lot of CEOs that I know just want to outsource everything to the cloud. Is your dean the same way? Like just outsource them. No, not at all. It's very difficult for us to outsource things, try to put things out in the public cloud because of all the regulatory requirements that we have and nobody feels comfortable with putting our data out in the cloud when we have things like HIPAA requirements to deal with and who's responsible for that data if something does happen to it, gets lost, stolen, somebody hacks into the system. So, traditionally we don't sort of fit into a lot of the molds of some other organizations fit into where they're just very willing to put things out there. We're very protective of our data and our PHI. Health information of patients is very near and dear to us and we like to hold on to it and we want to be in control of it. So, what we've done instead is tried to build sort of a private cloud and without even saying that we're building a private cloud, we've tried to do that with Repar and HP and VMware and put all those things together and really leverage the infrastructure that we have and be able to be very agile and adept at what we do with all the requirements that are being thrown on us all the time. Yeah, so a lot of people don't like that term cloud. A lot of IT people that I talk to sort of look down on that, right? Right, I really call it outsourcing. It's really outsourcing. Yeah, so the public cloud is a form of outsourcing. It sounds nicer then. But it's good marketing, right? IT people don't really usually market themselves. They're actually quite humble, even though they know they're solving some of the world's hardest problems. But we could talk generically about cloud, public cloud, private cloud, hybrid. Do you, so you're building your own sort of private cloud or IT as a service, is that sort of a direction that you're going toward? I mean, I think that resonates better with most IT people. Yeah, absolutely. We're trying to build a very, building block approach and be agile in the sense that when somebody comes at you with an application that they want to run, that it takes a very short amount of time to build the components for that application and get it up and running and stand it up. And so we're trying to do that quickly, efficiently, effectively and at a good cost point. So, three part uses this notion of utility. I mean, does that resonate with you? Do you guys that you're running your IT or storage like a utility? Absolutely. Just plug it in? Absolutely, we plug it in and it just works because a lot of people have been very skeptical about putting all those different workloads onto the three part. And we've had no problems, no service level problems with any of our customers. Like I said, we have a lot of different applications and we're looking to add more all the time. Is the skepticism because you don't have a lot of knobs and dials and bells and whistles to turn and tune? A little bit of is that, a little bit is the old way of thinking about storage and how it was provisioned in the past and the limitation of you have 10 spindles to work with, 10 discs to work with and being able to pool all that storage is a huge benefit for us in the three part model. So, people have been sort of the idea of showing me, show me how it works, show me, let's put a couple applications on there and let's see how that works and then okay, we'll add another application, let's see how that works and we've built it up over time and it's worked out really, really well for us. So, Eric, how long have you been in IT? Longer than I'd like to say. So, what's your favorite product ever? I mean, you could say, you know, could be an iPhone, could be, could be, I mean, what's the best product out there that you've ever had an experience with? Besides my Atari 2600? Oh, okay, that's cool. I love it. Wow, would you start in IT when you were 13 or? No, no, no, no. Everybody had an Atari 2600 here growing up, right? That's awesome. Besides that, I'd have to say right now, it's absolutely the three part. Absolutely, hands down. And maybe- That was not a setup question, right? I mean, we didn't talk about that beforehand. And maybe just slightly below that is our blades, our HP blades, so because- Really, you get excited about blades? I do actually, I love talking about this stuff because it works really well for us. It solved a lot of problems that the blades did and HP is very responsive to us as customers whereas some of the other organizations like IBM and Dell were not as responsive. They were responsive maybe in different areas. So I've talked to a lot of three part customers in my day. I get the same thing. Yeah, yeah, it's great, it works, we set it, we forget it. Is there something though that you wished they'd do better? I mean, what's on their to-do list? You know, I think like a lot of organizations integrating their software products with things like Exchange and VMware and SQL, I'd like to see a lot more development in those areas and making them work a lot better, making them work sort of in the three part model where there's not a lot of hand-holding you have to do. Absolutely is one of the things I would like to say. So application integration. Yes. Would be top on your list. Absolutely. Because that's what's driving the quote-unquote business for you guys, right? Absolutely. Yeah, so talk a little bit about that dynamic between the application folks, the infrastructure people, CIO, CIO, you know, a lot of CIOs I talk to, they care about risk. Right. You guys care about cost and efficiency. Yes. Application people, what do they care about? They just want their application to run well. They don't care about your cost, do they? They don't care about the cost at all. They just don't want to get a call from, you know, the customer, the end user saying, hey, this is a horrible experience. You know, I think that is where I come from sort of marrying the two things together, the back end infrastructure and the application as a service is, I want the same things. I mean, I don't want people calling me and saying, hey, you know, we're having problems with this. So, and that's the perfect scenario for three par because some of the things under the hood that it can do that if you are having, you know, maybe some areas where applications not performing as well as you thought it would with the setup that you put together is that you can change that on the fly. You know, you can optimize your storage, the way that you need it for a particular application. Like, you know, Oracle Rack, you're doing that or SQL or our cache database that we run for our clinical systems. What's the thing that you're most proud of thinking about specifically from the IT organization that you guys have accomplished in the last, you know, five, seven, 10 years? Two things, virtualization. I think we've done a really, really good job with virtualization. I think we've done a good job in explaining it and helping our customers understand what it is and what it can do. The density of virtualization as far as server-side virtualization, we've been able to get some really, really good densities using the HP, you know, C-class blades and using three-par storage. The other thing is this migration that we've done and we're still in the midst of doing it from our legacy email system to Exchange. But I think the three-par is definitely one of the things that, you know, I took a lot of heat for and it was one of those things that you can sort of kill your career. You know, if it doesn't work out. You stuck your neck out. I did. I did. And, you know, one of my coworkers helped me and stuck his neck out as well and we brought a lot of people over, you know, to the believers side of the equation and there was a lot of naysayers in the very beginning. They didn't know who three-par was. So it was unknown. It was the whole workload, mixed workload issue. Yeah, is this really gonna work for us? The magic juju of the three-par box, spreading data all over the place. Exactly. You know, I can see why people were concerned about it. I mean, you know, I mean, you lose your data. You, not only do you lose your job, but somebody else's head will roll. And if we lose our clinical database, you know, that's huge, you know. So those are things that, like I said, you know, can kill your career. But, you know, from a practical standpoint, it may take you a long time to recover from something like that. So that's not something that we can necessarily afford to do. But, you know, one of the big things was that, you know, the CIO definitely believed in what we could do and our abilities and our capabilities to allow us to go out there and do something like this with three-par and HP. What got you interested in three-par? Just read about it, or did they call on you, or did that go down? We had seen, I had seen a demo about a year and a half before we started really talking to them seriously about it. And the technology was just different at the time. I mean, the model was different. You know, the model wasn't, you know, you have X number of disks, build your array, groups, you know, present out of LUN. It was, you know, we have this big pool of storage of all these disks that are running. And, you know, yeah, if you need to change it, you know, you need to go from RAID 10 to RAID 5 or vice versa, yeah, just go ahead and do it. I mean, there's a few commands that need to be run, but it's very easy to do. And being able to do the mixed workloads where, you know, we started throwing three, four workloads at an EMC and you realize you have a lot of hot spots all of a sudden, you have IO problems, you know, and from the node perspective on the three-par side, it's much more scalable. So when you run out of capabilities, if you run out of cache, for instance, in EMC, you have to buy another array. You know, if you run out of cache, which I haven't seen in the three-par, you buy another node pair, right? You just build it out. So, you know, from the front-end standpoint, you have a lot more connectivity on the three-par side. You can connect hosts up in multiple channels, multiple pathways that you couldn't necessarily do with EMC. So you saw that, the light bulb went off, and you said, all right, I'll take a chance. Absolutely. And it's worked out pretty well. What do you make of this whole converged storage? I mean, you know, the three-par iBricks thing is there. And then you see, you know, more of a converged story around products that you, I don't think, have in your shop, like a left-hand and iBricks. Right. Is that something that you want to see three-par do? Do you see advantages of that convergence, or is that more sort of marketing to you? It's not marketing. I think, you know, with the organization, the way that we're built, and, you know, like I said, the amount of resources that we have, it really helps us to have a single platform, software platform, that we can actually manage that with as few people as possible. So the three-par platform really works well for us in that sense. You know, I would love for us to maybe get into some of the X9000 things like that, but it's not as practical for us. You know, there has to be a little bit more compelling reason for us to jump into yet another platform, which brings a whole lot of other things. So, you know, we're also front-ending our NetApp with three-par as well. So it gets complicated really quick. All right, Eric Krueger, we're here with Eric at HP Discover, an IT practitioner, talking about three-par, your adoption of three-par, the chance that you took, it worked out really well. I'm glad. It worked out. We've always been very impressed with three-par and was kind of a catalyst toward simplicity in the data center. So, well done. Thanks for coming on theCUBE. It was great to have you. Thanks for having me. Appreciate you sharing your story with us. Thank you.