 Live from Las Vegas, Nevada, it's theCUBE covering EMCworld 2015, brought to you by EMC, Brocade, and VCE. Welcome back to SiliconANGLE TVs, live continuous coverage from EMCworld 2015. Digging into the VCE, back into EMC as a part of the Federation Day here at the show. I'm Stu Miniman, joined by my co-host Steve Chambers. I'm really excited to have an IT practitioner here, David Sampson, who is the EVP at CTO, an I-trica corporation, company based in Quincy, Massachusetts, so not far from the Wikibon home-off space. David, thank you so much for joining us. Good afternoon, gentlemen. Thanks for having me today. Alright, so you've actually played with one of the technologies that's been causing a lot of buzz here at the show. So, VCE's main announcement this week was on the VX rack. You've been using the underlying technology scale I own now, my understanding is for over a year. About 18 months. So, first, can you give us just a little bit of background for those that don't know, what does I-trica do, and what's your role there? So, I-trica is a company that provides technology solutions, purpose-built based on customer specifications. We're about a 10-year-old company with a very strong history of working in regulated spaces, specifically the pharmaceutical space, but our experience in doing that has brought us outside of the software development space for compliant industries into producing infrastructure solutions. And we've been doing that in many verticals for about five years now. Yeah, so you created your own software, but now you're a full service provider. Right, so our software was really to facilitate clinical trial processes. Great, and, you know, one of the things I want to dig into you on this session is, what, as an IT department, do I do versus what do I go to someone else for? And there's some of that even for your company as to, you know, do you build the data centers, do you build the software, or do you there? So, can you sketch out for us just a rough sketch of what your IT department and, you know, IT resources look like? Sure, so from our perspective, you know, we're providing hosted solutions for customers primarily. Typically, virtual private cloud, private cloud, sometimes multi-tenant cloud, but that's not the typical model. We are talking to more customers today. We are seeing more customers ask for some on-prem solutions. Customers looking to spend some of those resources and workloads back into their internal data centers for cost control more than anything else. All right, so let's build your stack a little bit. You know, where's your stuff hosted? Do you build your own data centers and, you know, what infrastructure do you use inside? So we use outsourced data centers. Primarily tier 4 facilities. We're right at Supernap down the street here in Las Vegas. We're also at Napa the Americas in Miami. We have a tier 3 facility in Boston, but in core site. So we don't build our own data centers. We go to the people that are best in the industry for data centers. We don't think there's a lot of value in building data centers today because there's folks doing it on a very large scale and they're doing it very well, better than we can do it. So we go into those data centers on a really a co-low model. We implement the servers. We build out those server stacks, and we manage those server stacks to the customer needs. All right, so yeah, it's great what lead into, you know, somebody else can do the facilities, power and cooling, all that wonderful stuff. Physical security when it matters for stuff. And let's talk, what led you to Scale.io? So traditionally our company has offered some very high performance storage solutions based on legacy flash technology. The challenge we always had with that technology was that we were locked into specific networking platforms. We were locked into single node solutions, and they created some challenges. So we were looking for a solution that would give us high performance with multi-node redundancy and kind of take the handcuffs off from any specific hardware or networking platform. Software defined was the way to go. We went out to the marketplace about two years ago. We started looking for solutions to do that, and we landed with Scale.io around the end of 2013. All right, so that was after the acquisition, but were you an EMC customer before? We were, and we have VNXs, and we were previously an EMC customer, yes. Okay, and so this was obviously before the, you know, VX Rack existed. How did you build that Scale.io solution? Right, so that was a great project for us, and it's led us to some very interesting and unexpected places. So we really spent the last 18 months validating Scale.io inside our lab, coming up with, you know, a number of different test routines to drive the performance of that technology as far as we could. And we've deployed stacks with the Scale.io solution, all SSD based, spindle based. We've used a variety of super micro server platforms to achieve the performance that we have combined with Malinox networking technologies. And we really landed on that, you know, when folks talk about Scale.io, they talk about flexibility around networking. You can do InfiniBand, you can do 10 gig, and there's a lot of choices for that. A lot of our legacy infrastructure is operating in InfiniBand, so we were almost forced down a road of maintaining InfiniBand in our environment. Fortunately, through our partners at Malinox, with the InfiniBand Ethernet gateways, we've now been able to bridge our legacy networks over to Ethernet networks, and we're using 56 gig Ethernet as an end-to-end solution from our new storage infrastructure to newer hypervisor hardware. And we're a company that operates multiple hypervisor platforms. We do operate some bare metal, but typically all of our workloads are running on KBM and VMware. Yeah, it sounds like, I mean, that was a perfect tip for you going with the Scale.io then. So I want to talk a little bit about more just the role that hardware plays and really the testing. How long did it take you to build that solution, and what was involved in kind of making a stack that fit for your application needs. So we spent a good part of last year validating hardware. You know, extensive, extensive testing, we built for the fastest possible use case to replace that legacy flash hardware, and that meant that we invested the top end that we could. Again, we chose a super micro platform. Typically we were at HP shop previously. We moved away from that, got into a little more commodity hardware play, and we found that that's been very successful. We've had no issues with that hardware, so it's worked extremely well. Okay, cool. So how big is the Scale.io? Did it start small and grow big over time? So we started testing with four node stacks. We do run multiple environments for different customers. Some of those environments are running stacks up to 20 nodes today. We have a, you know, we have an assistance development lifecycle process. So we work with our customers to develop a specification of exactly what they're looking for in a business and technology solution. And as part of that discovery process, especially from a Scale.io perspective, we have a really about a 50 question interview process that we work with our customer so we can tune that environment to exactly what their needs are. We have extensive bench in Linux operating system. We've found opportunities in the Linux operating system to tune that to meet the needs of Scale.io based on our customer's needs. So depending on block size, depending on I.O. requirements, we can choose the hardware that fits their environment and then tune the operating system to their needs to get the best possible service delivery. Yeah, so I wonder if you can expand for us a little bit. You know, look at how much effort it took you to build that stack. You're working with VCE and the VXRAC. How much is VCE just going to be able to do that for other customers versus the flexibility? Because service providers are notorious for wanting to be able to turn all those dios and tweak it for their specific needs. I think the VXRAC provides a great opportunity for a lot of folks. You know, VCE has provided a great customer experience for their service, their customers, I'm sorry, for a very long time. And the ability for folks to be able to leverage Scale.io inside that stack is very significant. What I'm seeing, and you know, we've been here at the show for a couple days, I've talked to a lot of folks about Scale.io and it's been very exciting for us because we've validated a lot. But one of the barriers to entry for Scale.io is how do I do this? How do I deploy the solution? What's the hardware? Is there a validated reference architecture? How do I know what I'm going to get? So having the VXRAC available as a platform that, again, can support bare metal, multiple hypervisors, and is effectively a plug-in solution for VBlock and non-VBlock customers is really going to be a great tool for the marketplace to be able to provision very fast, less than 30 days. I know that we've all heard the VCE buzz, but it's the reality. It's what folks are looking for. So, you know, I heard you talk about customization for customers, you know, tuning and things like that. You know, aren't you going to lose that capability if you go to VXRAC? You know, don't you just get like a standard build? How do you reconcile that, you know, that VCU going to take away, you know, what are the hard work in validation? Totally get that. Are you losing anything, or is that not as important? So the VCE product is going to have storage heavy nodes and compute heavy nodes from a hardware perspective. So from that perspective, they have several profiles of hardware they'll be providing customers. So it's not completely rigid. It's a little bit... Well, from the operating system perspective, I honestly don't know. So, you know, my expectation is that, you know, standard server hardware, we're operating Linux OS, and we can go in there and turn the dial. I doubt that the hardware, the operating system, is hardened to that level, but that's a question for somebody else, unfortunately. So, David, you know, my understanding from, you know, we talked off camera was in addition to doing Scale.io, you also have VSAN. A lot of people out there, you know, are like, oh, I don't understand how the EMC Federation can have these competing products. You're using both. Can you give us a bit of an understanding as to where they fit in your environment and how you see them? Certainly. Give us your experience, not the EMC space. Yeah, so, you know, just over the beginning of last year, we were going down the hyperconverged road. VSAN was coming out to GA at the time, which we were very excited about, and we embraced that. We are running production environments with VSAN. VSAN is a great technology. The challenge with VSAN is that it comes as it is, and it operates based on what hardware you attach to it. Flash, SSD, and spindles. There's not a lot of knobs to turn there. And given our experience and our expertise, you know, we found some bottlenecks there that we worked very hard to optimize, but we're ultimately limited. And Scale.io is really the ultimate software-defined storage platform because the hood is completely open on the car. You can bolt on the pieces you want, you can make the operating system tweaks that you want, and you can make changes on every single level. It's a much more known quantity from what you're receiving from a storage deliverable than VSAN. It's interesting. So you've got different flavors of kind of similar technologies. Looking forward, how do you see your adoption going forward? Are you definitely going down the VX route? What about VSAN? Can you see yourself testing that again? What was the future look like? I think as VSAN evolves, we'll continue to keep an eye on it. It's certainly a very important technology. We are a VMware partner. We use VMware extensively. From the Scale.io perspective, Scale.io is what we're using today for production workloads. We've been operating it for since third quarter last year. We're moving all of our tier one workloads over to Scale.io based platforms. We see it as the ultimate storage platform. Its ability to operate under multiple hypervisors and might put multiple operating systems reduces our administration overhead. In addition, from a hardware perspective, we're saving about 50% on our cost of storage, which is very exciting. So David, I'm curious, what does using a solution like the Scale.io do for your ability to meet your customer's requirements, add new features, spin up new customers, how does it impact your business? So the flexibility of Scale.io has created great new opportunities for us. We've developed significant automation around the platform. So from a deployment perspective, one server hardware is deployed to a rack and connected to networking. The deployment process is fully automated from an operating system as well as an application perspective. So that integration has been tremendous for us, saves us considerable overhead, allows us to serve our customers faster, and then we've integrated further into monitoring and management capabilities as well as down to customer billing, because everybody cares about customer billing at the end of the day. I certainly do, but our customers do as well. It allows me to monetize to them the services they're receiving. So we're getting the plug, but I want to give you one last question, David, because you've been so early in working with EMC and with VCE so much of this. So what's next? What's on your wish list that you pushed to EMC, VCE, or the industry in general? What's going to make your job easier as a server provider? Is it scale? I mean, I know you're testing more than 100 nodes now. What's on your wish list to make your job easier? So seeing scale is very exciting for us, and we really are very excited about that. I know that the EMC team is working on enhancing the out-of-the-box scale feature set from a management perspective on a lot of different levels, and there's also talk about integrating some additional technologies, like maybe Ddupe or compression. We've heard some rumors about that, and those are all things that we can use to deliver extra value to our customers. There's no question from our perspective. Scale.io is the ultimate storage platform. Being able to get that inside the VXRAC technology is going to be a huge value for VCE customers, and I have no doubt that when the VXRAC product starts shipping, it's going to be very heavily embraced in the marketplace. All right, well, David Samson with I-Trika, really appreciate you sharing this. Appreciate what you're doing to push on the vendors to get these technologies moving faster. We've said for the last couple of years, service providers really are, in many ways, the new channel for these products. A lot of cool stuff going on there. I'm Stu Miniman with Steve Chambers. We'll be back with lots more coverage from EMC World 2015. Stay tuned.