 Live from San Francisco, California, it's theCUBE at VMworld 2014, brought to you by VMware. Cisco, EMC, HP, and Nutanix. Welcome back to San Francisco everybody. This is theCUBE and we're here at VMworld 2014 inside of Moscone. We are unpacking some discussions with practitioners here. We always like to talk to guys who are actually using the technology and helping drive business value. Mario Angers is here. He's the system manager at the University of British Columbia within the central IT group there. Mario, welcome to theCUBE. It's good to have you. Thanks very much for coming on. Thank you. So VMworld 2014, let's start there. I mean, a lot of action. Very exciting, yeah. A lot of new stuff, a lot of pretty exciting announcements. What I talk to practitioners, they say, I love to come to these shows. I get to talk to my peers, talk about what's working, what's not working, learn about new technologies, talk to the engineers. I presume you're doing a lot of the similar stuff. Yeah, yeah, it's a really, really good opportunity for networking. You get to talk to the people that are actually building the stuff, right? So you can provide feedback face to face. So yeah, it is, it's a fantastic opportunity. You're giving feedback. What kind of feedback you've been giving? Some positive, some, you know, this is, you know, certain things need to be fixed. They need to get better, right? So break it down for us. What are you, as a customer, VMworld customer, what are you, obviously happy with the, you know, the first wave of server virtualization, that was a no-brainer. So I think the strategy behind what we're trying to do is trying to decouple the technology from the actual service itself. So the technology, you shouldn't be reliant on a specific technology to provide your service, right? So yeah, VMware, you're kind of married to, right? There's not a lot of, you know, there's not a lot of competition in the marketplace, so there's not a lot of other folks you can look at to provide that, you know, that technology stack, but anything behind it, you know, the important part is having the ability to, you know, to be able to replace those components, compute, storage, network, right, without impacting your clients. So when you're talking about service, you're talking about fundamental services of capacity, storage capacity, or networking, bandwidth, or compute, or underlying services. To me, service is higher than that, right? It's essentially infrastructure as a service, or platform as a service, right? So DR, backup, right? Data protection, data movement. Yeah, things like that, right? Or email as a service, desktop as a service, right? You know, things like that. All right, okay, so you're saying that today there's a tight coupling between the hypervisor, essentially the platform. There can be, depending on how it's built, yeah. Okay, and that's a vendor issue? That's how it's... I think it's a little bit of both, right? Because, you know, of course, if you, you know, for example, you know, we've got a couple different vendors in terms of storage, right? So Coho being one of them, right? And we also have some NetApp, right? So historically, it's been extremely difficult to move workloads from one storage vendor to another, right? For one reason or another. They're sort of locked into that array. Yeah, yeah, to some degree, right? So, but, you know, some of the things that are being done now allow you to do that, to move workloads between those various different vendors, right? Without disrupting the end user or the client. Talk a little bit about your organization. I jumped right into that. So University of British Columbia is the largest in, I would say, Western Canada. So we have roughly about 60,000 students. And, you know, we have roughly about 15,000 staff, right? So we need to support those folks. So we've got a fairly large IT infrastructure to be able to support that. You know, we have roughly about eight petabytes of data, and we've got, you know, 4,000 virtual machines to support the community. And that's growing all the time. And one of the things that we're trying to do now is extend the services that we provide internally to our folks at UBC, to other universities within the province. Okay, now you guys do a lot of Oracle databases, right? Yes, we are. We're a big Oracle stuff. So talk about that a little bit. So in the early days, right? A lot of practitioners were scared about virtualizing Oracle because Oracle was saying, well, you might have to go back to a physical infrastructure if there's ever a problem. They'd get you in a headlock and- A little bit. And their licensing model is really not conducive to virtualization, right? But people said, you know what? There's value there. We're going to do it. Damn the torpedoes. We advise folks in the Wikibon community years ago to do it. The Oracle will support it. And I'd like to hear your thoughts there. So is that what you found? Absolutely. So we went down that road, I think, four years ago, about four years ago now, and we've been very, very successful at virtualizing Oracle. So all of our databases, with the exception of one legacy, are now virtualized, right? So when we run PeopleSoft, we run our student information system. We run several enterprise applications on Oracle that are now virtualized. Okay, and you're providing databases as a service, essentially, or? No, that's not something we provide today, but that's something we want to look at doing, yes. Now, how does the storage fit into that? You mentioned Coho data a second ago. How does the storage fit into that database as a service? So it's really building on what I was saying originally. So to me, you should be able to deploy technology behind the service, and it should really manage itself. So you shouldn't, if you need to, you're always going to need some level of specialization to be able to manage the technology behind the service. But if only one individual in your team is capable of managing the technology, then I think you've done something wrong, right? So and I think what those new vendors like Coho are doing, they're simplifying the process of managing storage and other infrastructure components that specialization that was required a few years ago to be able to manage that technology no longer exists. So as an example, I had two people full-time managing storage all the time, that's all they did. So today I have nobody full-time managing storage. Really? So the entire team can effectively manage storage. You essentially have an arm and a leg and a wrist managing storage. Okay, so. So it's largely set and forget. Okay, so tell me more, why is that? Tell me more about Coho. So the CTO of Coho is also a professor of computer science at University of British Columbia. So we've been working together for, I would say probably two or three years now. So a lot of the things that they do, they come to us to test, right? So they'll come up with a new feature, we'll implement it in our environment, we provide feedback. They change based on the feedback we provide. So it's been a great relationship for both UBC and also Coho. So how is it different than sort of a traditional storage? So I would say, so first of all, it's built with essentially off-the-shelf components. So you don't need, again, it scales very quickly, it deploys very quickly, it's just, it's very, very easy to manage. Compared to some of the other vendors, it's significantly easier to manage. Interface, click wars. Yeah, essentially it's plug and play largely, right? And in order to scale, it's really not much more complicated than that. It's plug and play, right? Okay, and you found these guys because you had the relationship with them, right? And now a lot of companies are nervous about dealing with a small company. How's that been, how's that experience been? So, and I understand why a lot of people are nervous, so you don't understand if they're going to be around for a long time and all that stuff. But from my perspective, I think a lot of the innovation that is coming out of the marketplace today is coming from those small companies, right? The startups, and if we don't foster that, then the innovation that is coming out of the marketplace is going to go away. It was interesting, we just had a CIO on the county of San Mateo, and you're hearing that more and more from organizations and it's not, it used to be, well, if you're in the university environment, you're budget constrained, you know, if you're in government, you're budget constrained, so you're going to be sort of a laggard in terms of technology adoption. We're hearing more and more that actually to drive cost out into improved service levels, organizations are taking sort of be more aggressive with regard to technology adoption. Do you agree with that? Is that a change? I absolutely agree with that. I absolutely agree. So people are taking more risk, so to speak, because they see the payoff as lower cost, better agility, being able to do some of the things that we talked about, for instance, decoupling the technology from the services. I want you to talk more about that. Talk about how you intend and what your plan is to get there. I think the plan for us to get there is really understanding what the clients are looking for and then really designing a service that will allow to deliver what they're looking for while looking at how we maintain a high level of availability without disrupting our client and the ability to shift the technology that we use to provide the workload. As an example, let's use backup as an example. So backup normally it's done behind the scene. People don't understand how it's done. It's just there when it's needed. Well, a lot of vendors, they will actually nail you to their technology in order to effectively do backup. So we took a different approach. So we went to a fairly new vendor in the marketplace, which is Veeam, which is now allowing us to use pretty much any storage as a target for backup. Previously, we were essentially married to one vendor for backup and now that's gone. And we're looking to do exactly the same thing with storage. The need for storage is changing also. So historically it was either you're storing unstructured data, database, blah, blah, blah, and high performance wasn't really a huge requirement. But now based on what is being done in the marketplace, like big data, some of those things, it's becoming very, very, it's becoming impossible not to provide high performance storage to your consumers, right? And a company like Coho provides that. They give you the option to essentially use the various different technology that they're providing. You mean like a flash, for example? Exactly, yeah. So okay, so how are you using flash? You can talk about that. Today we're using flash for some of our databases, like the high transactional database that require a huge amount of IOPS to and from the desk, right? So we'll put those on flash. It's providing better performance. It's providing not just better performance for the database itself, but for the end user as well. We're also offering it as a service to our consumers. So they have the option to select a high performance desk or a flash tier to host their applications on if they do choose to. Now what are you finding? How are they consuming it? Are they saying, okay, I'm going to put the whole database on that flash tier? No, because it's cost prohibitive, right? So what they're doing is they're really being strategic about where the data is placed, right? So if it does require the high performance, then it goes there. Otherwise, they're leaving it on the floor. How does the data get placed? In terms of? So somebody says, okay, I want this sort of data set on flash. Who does that? How does the service handle that? Is it dynamic? Is it quasi? Today, the application owner really decides where the data's going to reside with our help, of course. But you know, those guys understand and know their applications very, very well, right? So they're effectively deciding where the data's going to be placed. And then what? They come to you and ask for assistance. So it's a self-service model. Okay, and what enables that self-service model? So what we've done is we've used vCloud Director in order to enable that self-service model. But now we're moving to vCloud Automation Suite, right? To offer that same, essentially, but with enhanced capabilities. So that simplifies it further. Yeah, the experience will look a lot more now like an Amazon or a Microsoft as opposed to, you know, vCloud Director was an okay tool, but it was really a tool built by engineer for engineers, right? Not for end users. When you say, look more like Amazon, you're trying to create this consumer-like experience. Yeah, a shopping cart experience. Okay, swipe the credit card. Yeah, you select what you're going to want right through a dropdown and then, you know, it just provisions. Where are you in terms of being able to provide that sort of services menu? So we're planning to launch in the next two weeks. Really? Yeah. Oh, so you've been busy? Yeah, very busy, yeah. Now, talk about that project. Did you do it in little pieces or is it one big roll over? So what we're going to do is we're going to launch in September as a pilot, so we're going to have four other institutions outside the University of British Columbia use that service. We're going to do a pilot for a few months, then we'll use the feedback that the guys provide us from the other institutions, you know, to really, you know, fix what we're not doing very well. And our goal is to go production at the end of November, early December, right? And then the intent is we're going to be providing service to every other school in the province. Okay. And so the rollout sort of by the end of this year, you're going to be in a position to essentially go live with this thing broadly. Yes. Oh, interesting. We were talking earlier about sort of some of the complexities of storage. Where do you see that whole thing going? You see VMware doing vSAN, talking about VVOLs, you saw the thing they're doing with EVO rail. Where do you see that whole thing going? Well, I think what they're doing is they're really executing on what we're asking them to execute on. Right? Like we all want simpler infrastructure behind the service. Right? So, cause you're either spending money on technology or you're spending money on people. Right? So it's really a choice you're making. My preference would be to spend on both. But you know, well, and then what I can do is I can redirect the people that were historically working and spending a lot of time supporting infrastructure on a higher value task and project. So another way to say that is you will spend money on technology to save time. Yes. Absolutely. And reallocate that time to higher value activities. Yeah, then we become a lot more strategic. We can, you know, we can essentially provide what our customers are looking for. Which is kind of the opposite of like an Amazon, right? They'll spend time with PhDs. They're trying to figure out, okay, how can we take cost out of this infrastructure? You don't have those type of resources. So this is sort of an example of sort of that mentality of hyperscale coming into the enterprise with a prepackaged solution. Largely. That allows you to free up time so that you can put time elsewhere. Higher up the stack, applications. How do you spend that free time? That freed up time? Well, so we spend time talking to our customers. We spend time understanding what they're wanting us to provide. You know, we also spend time delivering on what they're expecting us to deliver. Right, as opposed to just, you know, we, historically what we would do is we'd provision a virtual machine and then need more CPU, need more memory. All of that had to come through us, right? So we became the bottleneck. Now we're enabling them to essentially do all this themselves. So take the, you said the Coho implementation allowed you to save essentially nearly two FTEs. There's still some arms and legs going on, managing storage, obviously. So let's say it's, I don't know, 1.75. FTEs that you saved. Times they're fully loaded cost, boom. That's the ROI, right? That calculation. And you were able to redeploy those individuals in terms of them spending time on aligning with the business essentially. They had the skill sets to do that. You didn't have to re-skill it. I mean, in some cases we have to retrain, which is fine, right? There's nothing wrong with that. Everybody wants a good challenge. Sure, right. But yeah, certainly this is really one of the problems. What we're trying to do, right? We're trying to give people careers, right? We're trying to get- Go pass. Yeah. Great story. All right, Mario, if we're out of time, thanks very much for coming to theCUBE. It was a pleasure meeting you. Appreciate your insights. All right, keep it right there, buddy. We'll be right back after this word. This is theCUBE. We're live from VMworld 2014. We'll be right back.