 Hi, this is Stu Miniman with wikibond.org. Here with SiliconANGLE TV's live continuous coverage from Dell Storage Forum in Boston. We're day three, getting towards wrapping up. Joining me for my co-host for this segment is John MacArthur from Walden Tech. And our special guest for this segment is Lars Veccheridis. Lars, welcome back to theCUBE. Thanks. So, we had to squeeze you in the schedule. You're really busy here. A lot of announcements going on Dell. And yet, you were part of the keynote this morning. So, we appreciate you finding the time to come in here. So, first of all, what's your experience been like this week? So, I think you kind of summed it up for me. It's been jam-packed with discussions with all sorts of customers and analysts. It's, you know, I really enjoy the feedback. It's actually really gratifying to actually see so many people really engaged with our products. On the other hand, it's kind of interesting. It's kind of drinking water from a fire hose. So, I'm going to probably take three days after this and try and digest what it is that I actually learned. But a lot of really good stuff. I'm really amazed at just the excitement that's been generated. Well, Lars, you know, you've been part of kind of the virtualization ecosystem. We find they're high-energy people. Can you just give us, for people that aren't familiar with you, just a quick synopsis as to kind of your experience coming in from Equalogic, you know, through the Dell acquisition and, you know, what you do today? Oh, okay. What it is that I do. Well, so I was actually one of the first engineers that came in at Equalogic. I was initially running the, what we called the embedded system management team. So, I was in charge of building pretty much everything that you touch and see inside of the Equalogic arrays. And my role has grown steadily over the years to touch virtualization outside of the arrays. You know, the virtualization specifically with Microsoft and VMware. Also, the application infrastructure and the application ecosystems. And a whole bunch of other, you know, related systems infrastructure projects. And so, right now, you know, after Dell acquired us, obviously everything just became supersized. I now am the executive director of software development running, you know, pretty much anything that is not associated with the platform, so to speak, of the Equalogic storage. So, you know, anything that's sort of tied to hardware or very specific to hardware is something that I don't do, but all of the, much of the firmware, the top half of the stack, and then all of the external software is, you know, I get blamed for everything that goes wrong and my smart guys get all the credit when everything is good. That's a good bandage. So, Laz, as you start, you're starting to incorporate solid state disk into some of your products now. Sure. What does that drive in terms of requirements on the software stack? Well, so, you know, for many, many years, we've had the benefit of sitting behind some really slow rotating rust, and so you have a lot of liberties that you can take. You can write really bad code. Not that we wrote bad code, that's absolutely not true, but over time, you know, it gets tuned to the cadence of those disks. So, once you start putting solid state disks and everything goes much faster, suddenly you realize that you've created bottlenecks and those get unraveled a couple of different ways. One, you could actually fix the software and we've done a fair amount of that. The other, of course, is to get faster compute and we're doing that as well. So, over time, you know, I think now we're pretty much in good shape to be all solid state all the time if that's what we really wanted to do. And, you know, I think pretty much everyone says that, you know, that's the future. And in terms of the interconnect technology directions on interconnects, what are you going to do? Well, so you know that we're 10 gig, we support DCB, which is, you know, a protocol in the ethernet that allows a more, not necessarily differentiated, but a more reliable, deterministic quality of service from your ethernet. And at 10 gig, we're now at the point where, you know, most of those, pretty much all of those pipes are not saturated anymore. So now we're waiting for everyone else to fill them up so that we can have a bandwidth problem. 40 gig is coming, as you know, it's down the road and clearly we're going to be involved in upgrading our systems at that point. And then we're going to have yet another set of system ball next that we're going to have to worry about with the size of the compute behind that. So it's an exciting time. So Lars, one of the big trends if you talk about what server virtualization has done is that we talk about, you know, is the infrastructure meaningful? Paul Maurits has said, we're trying to make it invisible. You know, Steve Herrod on the VMware side said, we're just trying to kind of automate it and automate it and make it simple. At your keynote this morning, you guys had an interesting new vision talking about your piece, which is the software component of the storage and creating, you called it, the host virtualization storage, virtualized storage, or for all those people that love TLA's, it's HVS. So can you give a little explanation? Sorry, we don't have the whiteboard here, but what does HVS and what does it mean to customers? Well, basically HVS, we started with the notion of taking the platform specific stuff out of the package altogether. What would you be able to do if you could just take the storage management and storage virtualization components that are in equal logic array and apply them to either a generic or virtualized platform. And that's really what HVS started out as is basically a platform-less platform that allows you the ultimate in flexibility in determining what your running environment is gonna be. And so you don't have to worry about the types of disks or whether they're disks at all. They could be memory, they could be clouds, they could be anything that when you write something and you try and read it back, you get what you wrote. And so this is kind of HVS in its elementary stages. What's transpired over the last few years is that the cloud world has sort of blossomed and you have a lot more capable high density compute farms out there. Our new 12G servers, for example, are very interesting in these types of compute environments. You're gonna see very, very dense compute farms. And if we take this HVS thing and apply it there, you can get sand storage or sand storage management capabilities inside of a physical environment that doesn't look anything like a sand. So the concept that I presented today specifically was the notion that you could take HVS and carve little slices out of a bunch of these highly dense server compute farms and build yourself a group of ecological arrays and then expose these directly attached storage to applications through Iskasi and it would look just like an ecological array. This actually provides the service provider with a huge amount of flexibility because now you can scale the compute, you can scale the storage, you can move it around, you can add flash if you want, you can do all sorts of things. And as new technologies come in, it's very easy to just stick them underneath a virtualized array and take advantage of them immediately. Clearly there's a lot of software work that still needs to be done in order to fully realize that vision. We would have to fix our tiering and do all sorts of other things, but that's stuff that we're doing anyway. And so this is something that I think really resonated with me when I first heard about it and we finally outed it this morning. So you're targeting this announcement really towards the sort of cloud infrastructure application as the service providers or infrastructure as the service providers? There are a number of use cases in various places. So today this morning, the discussion was really, really tailored towards that. So infrastructure as a service providers or specifically there's a brand of provider that has a very homogenous infrastructure. It's just a rack of highly dense commodity servers and a switch on top or some variation, but everything is exactly the same. And you see a lot of those out there and in those environments, they need to be able to service applications that really want to talk scuzzy. And those aren't just, by the way, service providers. There are actually some very large enterprises out there that are really quite enamored of this particular type of data center architecture. But again, everyone runs the same applications and all the same applications still expect block devices underneath. And so this is, you know, I always thought ecological was the greatest way to manage block anyway. Yeah, so last, you know, we've had the number to talk to a number of your customers and that they really spoke well to kind of the scale out and simplicity from the first generation of ecological to the most recent, you can put them all together, kind of automatically put them in, take them out. When you talk about scale out architectures, one of the new areas that, you know, obviously we're spending a lot of time on, actually as soon as we wrap up the show here at Dell, we're going to be throwing to our West Coast broadcast, which is at Hadoop Summit, is big data. So do you have any commentary on what you see, kind of the Hadoop, the MPP appliances and the impact on storage? Because at a high level scale out is one of the big pieces. Well, you hit it, you know, exactly right on there. So I've actually wandered around the big data world quite a bit recently and looking at a lot of those projects. You know, and Hadoop is one thing. I think a lot of the intellectuals that I've discussed Hadoop with are sometimes not especially enamored of it. I mean, it does something very well and I think a lot of people are using it for that particular use case. But the key to having big data, especially if it's unstructured data, is having really, really large, contiguous chunks of data. And, you know, Hadoop, as you know, is the notion that you, you know, underlying it is the notion that you cut everything up into little chunks. So how do you reconcile the two? And scale out, as you say, is a really interesting concept. If you really need to have a gigantic, you know, tera-scaled data set, you need to have, it's much easier to manage it from a data analysis standpoint if you only needed to access it from one protocol endpoint and you could have a whole bunch of compute going in through one either logical or physical endpoint to access all of that data. Otherwise, you create a whole layer of complexity at the next layer up in your system. So things like scale out file, things like scale out block, you know, they're necessary building blocks, I believe, ultimately, in what big data will entail. I should note, though, that I'm not that good a prognosticator, you know, and this stuff, if you talk to the people that are doing it, they're the first to tell you that it's, this is, we're like version 0.5 of what big data really is gonna ultimately become. I mean, at the top of the stack is really the analytics and everyone is really enamored of what you get out of it. And I'll admit it, it's really sexy stuff. And, but the infrastructure is still in the process of congealing. So, you know, rather than trying to make a definitive statement on something that seems to be changing by the minute. For me, yesterday was about big insights, not big data. So, so even small companies can make big insights out of small amounts of data. Properly analyzed, so. So, so last, you know, as we're getting towards the end of the week, we've heard about, you know, all of Dell's acquisitions, how they're building the portfolio, how they're building blocks that all kind of weave together. Equalogic was really the first big piece there. So, from the development standpoint, can you give us a little bit of insight as what the engineers, how they work with all of these new acquisitions, how they integrate with the workforce that's now quite dispersed. You got the Minnesota guys, the Texas guys, the Silicon Valley guys. Yeah, so, yeah, so JetFuel solves all sorts of problems. And in this particular case, it does. I think Dell has a very, you know, sort of face-to-face culture in general. This is true for any organization that's always been together, but Dell is particularly interesting as you think about it. They haven't really been geographically dispersed. I mean, the center of gravity has been so predominantly. Pretty much around, Rod. Pretty much around, Rod. And so, when we were the only outpost, it was kind of interesting. It was a lonely existence. But, you know, I think as we brought these other companies along, we try and meet on, you know, face-to-face as much as possible on a regular basis. It's really important to get to know everyone. And then, the other thing that we've done, which I think we're doing really well, Carter alluded a little bit to it, but I don't think it was really more a subtle illusion than anything else. But, you know, we really want to be able to build stuff together, where we're leveraging each other's knowledge and skills and maybe even the actual source code in some cases, so we don't have to build something twice. And we've really worked very hard at doing that. It's a painstaking process because you have to get everyone in a room and they have to hammer out problems and they get a hammer out. There are particular versions of the problem, so you have to avoid boiling the ocean. You have to scope things down to something that's achievable. But, you know, the other thing that was discussed today, which was the RNA cache, is an example of everyone working together to build an architecture that's compatible with the entire portfolio. The one thing I will say, it takes a little longer, right? So, you know, the RNA acquisition was a year ago. And in that particular case, the company didn't really have a shipping product. So, it's a little while before you see the fruits of this labor. But, you know, we're really, really adamant about making sure that, you know, we're not building silos. You know, I should point out, there is one, the first thing we did was the X-net acquisition. You know, that integration has been shipping now with Equalogic for well over a year. And, you know, so it's kind of the same thing. Fly to Israel, meet a lot of guys, figure out how this whole thing will work, you know, write code, figure out, you know, why it doesn't really work. You know, you go back and forth a lot and eventually you get to a product that's really much more tightly integrated. All right, well, Laz, your next meeting's calling. We're getting the hook. So, we really appreciate you finding the time to come here. Congratulations on the announcements this week. Tremendous progress. We're definitely going to be watching closely the new stuff you're demoing and where things are going forward. So, welcome back anytime. Hopefully we'll see you at VMworld. Yes, absolutely. And some of the other shows. So, we will be right back with Silicon Angles live continuous coverage from Dell Storage Form right after the short break.