 Live from the Mendelebe Convention Center in Las Vegas. It's theCUBE, covering VMworld 2016. Brought to you by VMware and its ecosystem sponsors. Now, here are your hosts, Stu Miniman and John Walls. And welcome back here on theCUBE as we continue our coverage of VMworld in Las Vegas. For the Mendelebe, we're sure you can see how great activity we're going on here. We're going to be here bringing you coverage here Monday, Tuesday, Wednesday, this week. Along with Stu Miniman, I'm John Walls. We're joined now by Mike Field, who is the Interim CTO of Bay State Health. Mike, thanks for being with us here on theCUBE. We appreciate your time. My pleasure, my pleasure. Let's sit at the table here about Bay State. Western Massachusetts-based health care provider. But a lot of hats you're wearing is the Interim CTO. Teaching hospital, community hospital, home care, hospice care, 12,000 folks working in your organization, medical associates. How are you juggling all that today in the IT world? Difficultly, but it's actually not that bad. It's a pretty straightforward job. Love it. It's just new stuff all the time. I'm learning all the time. I'm working. I'm having a great time. We've talked a lot this week and heard a lot this week about the digital transformation, how very real that is, and moving on to the digital economy, if you will. Very important in your space in terms of patient care, patient information, patient data. From a 30,000-foot level, what concerns do you have in terms of that record protection and then that information's being sent about various diagnoses that you have to be concerned about? Sure. Well, first of all, we have to maintain up time. I mean, at the end of the day, our doctors, our clinical staff have to have access to all the tools, all the systems all the time. So we simply have to have everything operational all the time. And in today's budgetary constraints, like our institution, like many others, is going through a lot of budgetary issues. So what keeps me up, other than security, which is a whole other issue, is how do I maintain the quality of our IT, which is the foundation of much of the systems that patients rely on? That's it. That's one of my major focuses. Mike, can you sketch out for us just kind of the scope of the responsibility that you have, how many locations, the number of kind of patients or employees, however you measure that? Sure, we're about 12,000 employees. We had well over a million patients in our area. We're about three, really four hospitals. We have an insurance company. We have a reference laboratory. We have a life sciences. We're a vertically integrated health system, teaching hospitals, as you mentioned. My responsibilities are to help chart the technology direction of the institution as a whole with certain specific mandates. One, I have to increase uptime. I have to increase efficiency. I have to work within essentially declining budgets. I have to integrate new technologies, all of the above. All right, so talk to us a little bit about what you've been doing the last couple of years from a technology standpoint. I think most of our audience will hear what you said about the challenges and it's like, yeah, that's it. Declining budgets, flat to down on headcount. More things coming in and some of them understand the governance and compliance things that you also need to deal with. And I live in Massachusetts. I hear all the things going on in healthcare there. Well, technologically, I would say the centerpiece is hyperconvergence. To us, there really is no other way to control costs, increase performance, increase operations, everything. And that's been the major focus at the most fundamental level of our infrastructure. We began about two years ago, what we call Infrastructure 2.0, which was essentially the replacement of all of our traditional storage, compute and networking elements, all based on VMware's NSX, ESX, and VSAN products. We are now actually in production and we've seen phenomenal results from this. All right, so Mike, tell us, what was that replacing? And this wasn't a project deployment? You were trying to kind of sweep the floor? Yes, to more. Well, actually it's a good question. You really can't, it's interesting. We come at this two ways. One way is to say that there's an event like you put in CERN or Epic and that's a natural point where you can, as you say, sweep the floor. We didn't really have that as a case. What we had was our normal budgets, which by the way, all of this work I'm mentioning was done within the existing budgets. We didn't really require new dollars for this. But what we had to do, knowing the new systems coming on, the new imaging platforms, all of these things, required that we build our new platforms to take care of those things at current or less dollars. So we started with initial deployment, very small, side by side. Our typical infrastructure looked like any other hospital. We had, for example, EMC storage arrays, IBM storage arrays. We had a mainframe. We had AIX equipment. Lynn, every complete salad bar approach to IT basically. And then we began an active, active, active network. Three separate sites which combined storage, networking and compute and have been increasing that while decreasing the other stuff and migrating to it. It's an ongoing process. All right, so two years ago, virtual sand was a relatively young product compared to where it is today. It's had a huge growth, especially the last 12 months. So what led you to that? And I hate to say it sounds a little risky, going out to, I mean, VMware's trusted. It sounds like you've used them for a while, but why were you willing to be kind of an early adopter of that technology? Well, if you want to make a significant change, you're going to have to take some risks. Typically, healthcare is somewhat risk averse, but Bay State is kind of blessed. Our CIO, a gentleman named Joe Venko, is very forward-looking and realized that to make these changes, we had to make some very serious bets on technology. VMware, we're familiar with VMware. Everybody uses it. It's a solid company. Prior to my work at Bay State, I worked in other, as a CTO and other relationships and other hospitals. And VMware, once again, over there was well known. So we knew that VSAN had potential. We also knew, to your point, it was a new product. So when we started this project, we didn't have, we didn't plan to migrate any of our very significant loads to it until the product progressed in its life cycle. We were intimately in communication with VMware about what their plans for and when those features would come on and we mapped our project to that. So you talked about results. Yes. Get into this a little more specifically but the before and the after, what are your metrics, your key metrics, and how much did you improve on? Sure. Well, I guess one of the simplest things and most obvious is just cost of storage. When we look at the true cost of storage in our old model, which are the big sands, the fiber channel arrays, all the special people to do that, all the special software and what we have now combined everything really under VSAN, our storage costs are about 30 to 50% less on any unit measurement like on an IOP basis, on a capacity basis and so on. Another enormous benefit was that we had like a lot of hospitals silos. We had the little storage guys here, the networking guys here, the compute guys there. Each one of those groups was understaffed, highly stressed. When we put in the hyper-converged platform, VSAN, NSX really being the glue of it, they all started to learn each other's roles. So the net result was we had a larger pool of people to learn communal skills on call pagers where it need isn't much, we have a larger pool of people to handle things. In addition, as they got better and better at it, they slimmed the need of like real-time attention. They could kind of work on other projects. So we had a backlog of many other things that the IT department simply couldn't get to because of keeping lights on. This technology allows me to not only have everything cost less, but I can deploy my resources outside of just keeping the lights on. So that big benefit. There's like an automation efficiency then. Absolutely. You're offloading people from having their eyes up here and they can be up here and doing other things. Absolutely. Can you walk us through some of those operational issues? There's kind of an inherent fear sometimes, if I own this domain, this guy hands over the hair, I have our swim lanes, we understand the back and forth. You put us all in a pool, it might be easier to kind of cut. Sounds like everybody's got more than enough work to keep them busy. So it can help walk through some of those. Well, first of all, I want to step out and say that when we propose this, your point, the fear right up front, it's natural for you to say, oh my God, this is a technology that I could be replaced. If I don't measure up, I'll be out. That just didn't happen. Most people really want to learn. The problem is they don't have an option. There's no space within the organization's daily operations for which to learn something new. So after initial fear and kind of trust me, you're not going to get thrown out of here. And they began to see the interaction. They were involved with VMware at an engineering level. We brought VMware teams in along with my company, as I said, I'm a contract employee. And we created a training regimen where we pushed the stores and networking and the PC guys all into the same group. They worked through the actual design itself. We didn't, and VMware didn't just simply present the design. It was a requirement that our staff actually be part of the design process. So they learned this thing. Once that began, no fear. Everybody jumped in. All right, so Mike, you've lived through a lot of the maturity the last two years. Can you tell us what was it that was in the product when you said, okay, it can go in production? And what things are you still asking VMware for to kind of bring it even further? Sure, the initial, one of the things we always wanted was the ability to put outside elements into the VSAN environment for storage. So compute nodes that are completely outside the infrastructure need to be able to see the storage. That's something that we know is coming. Once VSAN was deployable in a three node architecture, and that's what we are. We are three separate, almost exact copies with dark fiber between them. So we can lose almost any conceivable combination of systems and still stay up. We needed that capability. VMware worked directly with us to give us essentially a pass as soon as their products were able to do it, even if it wasn't generally stated, we were willing to do that. So that was number one. Number two, your question about what we want them to do, a couple of things. Their all flash system represents right now some pricing elements. We're looking for, we have a lot of customers on VSAN that have spinning disk and flash. But to take advantage of some of the new things like erasure coding, certain compression, this all flash issue represents a potential capital increase. That's something that we need to work out with VMware. And VMware's aware of it. One of the issues is encryption. We're expecting encryption to be available. That's something we know that VMware's working on. Basically, most of our designs are in there and our requests are in their life cycle. So we're confident they're going to get to it. Are there areas, and this is almost a loaded question, it's inherent in your business that there are areas that you need to be super sensitive about. Talk about imaging, for example. So what are those that create for you, you think, distinctively interesting challenges than others in your position? Well, imaging, to us imaging means growth. The real issue of imaging and related parts of it is the fact that it's an explosive growth and the growth and the quantity of storage. Hospitals like Bay State are constantly exploring new forms of what we call modalities, X-ray machines, things like that. And these devices have ever more complex requirements. We really can't easily predict where storage goes. So I need a method where I have linear costs to storage. Right now, in the traditional model, with let's say EMC stuff, I have a certain number of drives in a case and I go to number of case two, then I have to add an engine and so on. And there's this stair step cost. I reach some maximum number and then I have to do the whole thing all over again. What makes VSAN essential to us, and it is a major issue in capacity, is that that's a linear curve. I simply buy more simple nodes. I keep throwing node after node after node into the system. It allows me to, if not predict where my storage is going in terms of quantity, I know where my costs will go because I can manage that. So that's what keeps me up. That's from this particular product's point of view. The security is another issue, which as I said, we know that VMware will be introducing. All right, so Mike, with the benefit of hindsight, if you talk to your peers, what would you tell them kind of about virtual, virtual SAN in general, hyper-converged, you know, overall, you know, that they might learn from what you've gone through? Well, it's definitely worth it. It's a little scary. I think time's on my side though. I mean, vendors like VMware, Cisco, all of them are pushing towards this direction. What I would say though is be very, very careful about how you deploy in testing. What we found is hyper-convergence requires so many intricate parts working together that as versions change, as VMware makes a new version of VSAN, and Cisco makes a new version of a, for example, Cisco shop of a UCS product, and then NSX has another version, these versions need to be tested very, very, very carefully. If you don't, you gotta, you gotta intend the consequences. One of the, not weak points, but one of the results of a highly compressed, hyper-converged architecture is that an error goes everywhere all at once. It's not isolated to a corner somewhere. So you would better test carefully. Teaching people to properly test model things is essential. I would argue that that, if you're not willing to do that, be careful. Sage advice from a man who has been there and done that, and is doing that in healthcare. Mike Field, thanks for being with us. My pleasure, guys. TheCube continues here from VMworld in just a moment.