 Okay, we're back. This is Dave Vellante, and I'm here with Stu Miniman. We're with Wikibon.org, and this is SiliconANGLE.tv's continuous coverage. We're here live at the Dell Storage Forum with Darren Thomas, who's the vice president and general manager of the Dell Storage Business Unit. Darren, welcome back to theCUBE. Thank you. It's great to be here again this year at Dell Storage Forum 2012. Yeah, we've had you on a couple of times last year, actually, as you might recall, down at Disney World. I love you. I didn't realize this year, you were born in Boston. Yeah, I parked my car here. You didn't spend much time here, did you? No. Well, anyway, you just came off the big keynote packed to audience. You were here yesterday introducing your boss. And good vibe here, really, I think a step up from last year. How do you feel? Well, I feel great. We're in our third year, really, of the integration work, and it takes a while to get there. So this is the year I've been waiting for, when a lot of the integration proof points are coming out. So I think you're seeing them in the announcements Brad made yesterday, the ones I made today. A great, great set of integrated products that show best in class times best in class. And it puts them in a class by themselves. Yeah, we've been talking a lot about integration, and we've made the observation that Dell, kind of unique to other storage companies, focuses on integration first. I guess in part you started with the blank sheet of paper, so that maybe made it a little easier, but not necessarily. You guys had to really focus on that, and that's a main tenet of your engineering focus, isn't it? Yeah, starting with the blank sheet of paper was very helpful, although it was a dead standstill start. So it had both good and bad. We had to start from scratch, but the beautiful thing is we didn't start with a bunch of legacy. We got to basically search the industry for the best in class. And that was really a design concept that was untested before. If you remember the kind of companies, storage companies were buying before we bought Ecologic, they were buying from, pardon my French, they were buying from the dumpster. They were buying companies that were in trouble. They were companies that were structured poorly. And we did exactly the opposite. We went for best in the world. And a lot of our resellers are here today. It's a very common theme to come up and say, hey, I was a reseller of Ecologic, you bought them. Then I became a reseller of Compellant, you bought them. Then I became a reseller of Ocarina or Apocher, and you bought them. And so my typical answer is, yeah, we buy the best. So you must be looking at the best. At what point did you guys decide in your history, your corporate history, to actually really go for it in storage? Did you sort of put your toe in the water and experiment, or had you decided at that point that you were going to be a major storage player? Well, I came to Dell in 2003 and in 2003 the company was pretty committed to the EMC relationship. And we gave that a try. The reality was the product set wasn't what our customers were telling us they wanted. It was very much great products but stuck on the legacy, if you will, not starting from the blank sheet of paper. And so very early, I mean probably by the end of 2004, I was starting to have very quiet conversations with the leadership at Dell about this revolution that we're on today. And I will tell you, Michael is an innovator himself. And so it took me about like three nanoseconds to convince him. So with a leader like Michael, it's not hard to get on a revolution. So you've got actually quite a large install base of Clarion products and what you're calling legacy products. How are you transitioning those guys to this fluid data architecture? Well, customers tend to transition at their own rate. A lot of them have projects they'll buy all along the year. And so when they come out with new projects, we typically try to transition them at that point in time. It's not hard. Customers can see the differentiation between a legacy system that's bound by the sheet metal it sits in versus a non-legacy system that's kind of virtual beyond its physical limits. And the customers get it. I'll tell you when they get it, usually when we show them a demo. They're both compelling and equal logic are the big seller. Same thing with Aperture. If we can get a customer to see a demo, it's a 99% chance that we're going to close. So I think them seeing it, because I can remember the early days of equal logic, I would explain it to a room full of customers and they'd be looking at their blackberries and they're doing the blackberry prayer, you know. And then you'd look up and you'd say something, you know, breakthrough like, you know, the licensing is free. And all of a sudden if they're a CIO, they'd look up, put their blackberry down and say, say that again. And that's when you had them. So it's not that hard when your products are this differentiated. Well, it's free and it's perpetual. And perpetual. Compellent uses perpetual because they charge you a fee up to a certain point, but then after that it becomes perpetual, you own it. And perpetual is a little different because you own it across the hardware. So if you change out the hardware, you can put your old license on the new hardware. You still own it. And so that's very differentiated. Everybody else in the storage industry, they sell you a new piece of hardware, you owe them new licensing. Yeah. So that question came up in the audience this morning. And that is different. You really don't see that. But at the same time, you guys made a lot of margin on that. Yes. On that software drag. Yeah. Yeah, no we do. And you know, for us again, it's good to be coming from a blank piece of paper. So for us it's all new. The other thing is the customers stay in a maintenance relationship with us. They buy the maintenance contracts. It's what they want. It's what we want. And so to be quite honest, we don't need to be greedy about selling them new software licenses each time. The customer is paying us a maintenance fee for the software license and we can continue that process from generation to generation. So Darren, if you look at pulling lots of pieces together and lots of different products, hardware and software, the management is usually a challenge. If you look at the unified space which you guys now have with the Exynet product and converged infrastructure with VSTART, you know, Dell pulled the management story together. Yeah, that is the holy grail, if you will, of the IT industry is all these management tools. And to be honest, the innovation rate is still so fast that you can't make the overall management device. A single pane of glass does not exist. Absolutely. It does not exist across all IT. And as far as I know, until it really, really commoditizes across everything, networking servers and storage, it never will. Because each one of those groups is, I mean, I come out with like 300 launches a year, servers comes out with, you know, 10 or 20, networking comes out with, I don't know how many. You just couldn't change the GUI fast enough. So the model is to raise it to the point where every one of these devices has an API and you use a provision, a policy-driven machine to provision the machine. And that way, the customer, you set it up with these individual element managers, but you manage it at a provision level. And we have projects underway to do exactly that. And where we can, we make compelence management interface, manage the file system. That we can do. But across all the systems, we're going to use a provision-level manager. And I think that's the converged infrastructure. That's the converged tool of the future. This is where I think the blank sheet of paper actually has helped you a lot. Because 10 years ago, when you talked about APIs with less modern development tools, it was like, oh, you know, not really a comfortable scenario. And it took a long time. With today's rapid application development, new tools, not only Java, but now you're seeing Node.js and other modernization going on. It just makes it a lot more of a viable strategy, doesn't it? Yes, it does. And the industry is moving to where customers want us to solve the bigger solution for them. So in doing that bigger solution solve, it requires this integration at this higher level. And I want to point out one of the interesting things about Dell. I know we're kind of a latecomer to this enterprise class, but one of the beautiful things about us is the server guy, the networking guy, and the storage guy. We all work for Brad Anderson. We all work for that guy you met yesterday. And the beauty of that is we're in meetings three times a week. So how networking affects storage, how storage affects servers, and all that connection is very tightly coupled. And I know some people came to the analyst meeting we had where I had Forrest and Dario and I and Brad on the stage. You got to see three of the four of us today. Forrest is on an anniversary love boat, so we didn't bother him. But if you see all of us acting together, you realize how close we really are. This is not... I don't like living a city a thousand miles away from these guys. We don't have to force communication. We work for the same guy. We're in meetings all the time together. And as a result, the convergence of Dell is not that hard to do. So Darren, one of the things I've been impressed with Dell is when they've made an acquisition, they're expanding where they are. So we saw up in Nashua, you've got a group you said over 500 people. In Minnesota, the compelling guys are growing there. We had Dario on yesterday talking about the growth in Silicon Valley. So it's great to see at the top that the management is working together. Can you give a little insight as to how are your engineers collaborating and working together across the globe? That's actually a great question because to be honest, you know how engineers are, they live in their own little worlds, and then you bring them together and they all fight very passionately for what they believe in. So we have an architectural meeting. You'd like to think this is among about five or six people, but there's about 25 architects. And we bring these guys together and nobody can referee their group of me. I'm the only one they all report to. So I go to this and it's like a day and a half, once a month, where we bring them all together and we ask the hard question, how are you going to sit this software inside of this solution? And there's a bunch of churning, but Carter, you met him yesterday, Carter helps us with that. We have our CTO Steve Looning and I sit as a triumvirate over this council of super smart guys and we write down, we're not leaving here until we get this answer. And so they get together at that architectural level. Once you've got the architectural consensus, those architects then return to their individual teams and they carry the message out because they're the leaders of those teams technically. So we manage the minds of the leaders if you will, we get them to a consensus and then it just happens after that. Carter came from Ocarina and of course you guys talked about putting compression into block today, which is not trivial. You made us a future statement of direction there. But you also got, you mentioned that Appashore has inline block and DDoop. Is that Ocarina technology? Initially, because we just bought the company four months ago, it was not, it was of their own design but we are integrating Ocarina so we will move to that. Ocarina model very quickly and the reason is, I mean we could have used both, we own the IP of both, but the reality is the customer benefit of having one compression algorithm floating through all the different solutions, we don't want a customer to try to have to remember how he has to rehydrate the data. If you did it on Dell, it's going to rehydrate one way. So you're going to make those hard choices and basically whatever with that other IP. Yeah and when, every time we buy a company, we're open with the company, we're saying, oh you did that, well we will probably change that technology with this technology because of our fluid data architecture do you have a problem with that? We give them, if they have a problem with it the deal's not going to happen. So I want to talk a little bit more about the business model because I'm intrigued by this. I said last year when I got exposed to the portfolio I said Dell can be very dangerous and the reason I said that is because Dell is a company that's used to PC margins and now you're in the storage business selling, you know, storage margins and you're not sharing those margins anymore with a major OEM supplier or reseller relationship and I know you like to talk about how it's all about the customer but there's a residual effect. You know, the guys in Broad Street love this higher margins in storage. So Dell doesn't have to live off of traditional enterprise storage margins. Everything north of PC margins lifts the tide of Dell, doesn't it? Yes it does. You know when I first came to Dell I was warned that if you got your margins too high you were supposed to reduce them cut the cost then go into the market and kind of drain the swamp that was the old strategy and I told them I said I'm going to flunk this math test because I believe that if I give you more money you're actually going to be happier and like I said it took about a nanosecond for Michael to get it and we're reinvesting that money if you look very closely we're taking that money and that's how we're growing Equalogic that's how we're growing Compellent that's how we're growing these so we're in the investment stage still so I'm putting more into Equalogic and Compellent than we're taking out and as a result those businesses continue to grow we grow the scale we grow this integrated value and so yes there's higher margins but we still like to think that we're turning those into more features faster and sooner to give our customers more value and you're in a dogfight I mean there's a lot of big competitors obviously that you're up against and that's why you've obviously been so aggressive in buying what you've called these best of breed technologies should we think of you as a company that's going to sell you've talked a lot about end-to-end so essentially you guys are selling into people that are doing business with Dell as opposed to necessarily going after discrete storage opportunities is that the right way to think about you or is it both? that's maybe the old way to think about us is that if you will we had this huge client business just about every company on earth we had a pretty big server business number one or two in the world everywhere and we could just follow those guys in the doors but typically that's a different buyer that is not the same buyer who buys storage or networking so Dario and I are our sales teams the sales teams that support us are actually going out looking and ringing doorbells of storage opportunities that may not be so I would suggest that probably 30-40% of my customers are new customers to servers and maybe not so new to clients everywhere but they may be new to the enterprise and so we don't just harvest if you will we don't just plow our old fields we're plowing new fields how do you decide when and where and how much to sort of drain the swamp and go hard at any price versus the value cell? it's a portfolio play we have a lot of customers who recognize the value and it's a great value at a great price we can beat the competition with our value stream and every once in a while you have to buy your way into an acquisition account or one of the competitors is just going dirt low and you have to follow them so it's more of a portfolio game our sales team manages this they have targets that they shoot for so it's in the construct of how a quarter is run by our presidents like Steve Felice and Michael manages that so I think that's managed very well but my preference is let's go after everything I just want my unfair share of everything well I mean when you're dragging along a lot of software and non-perpetual licenses and services you can more easily justify any price and go in dirt low you guys have a little different philosophy in that you're the new kid in the block so I want to talk about this notion of horizontal tiering that you brought up in the keynote what can you tell us about that? well horizontal tiering is kind of if you will the next step it's the fluid day to promise and it's very easy to tier within a box now you have to tier across boxes and it's not just tiering tiering implies that the array controller has control of that data sometimes I'm migrating the data I'm actually moving the data and even the servers know the data has moved that's in theory that's like an archive or a backup or data movement so we can do the data movement however we choose to do it whether we do it under the original controller or if we actually move the data Appisher moves the data and the array controller and the server and every part of the system knows the data has moved so in theory what we're going to do is if it's an equilogic device copying to another equilogic device we call that replication equilogic will manage that from the array we will be able to have equilogic move the data to a non equilogic array but Appisher will do it in a very clean way where the customer will get the benefit of the new server and the new device knowing the data exists there now and so there's a lot of data movement but we're absolutely going into this path where I'd like to think within two or three years we'll be able to move any data from any device we own to any other device we own at the customer's wish So let's talk a little bit about Flash you actually directly participated in one of the greatest value creations I've ever seen in the storage business between whether it was equilogic a left hand or three par, a compelling you saw Icelon you know multi hundreds of millions or even billions of dollars of acquisitions and now you see these flash companies getting very high valuations and many of them haven't even shipped a product yet I guess two questions there one do you think we're going to see a repeat of some of that frothiness that we saw with the virtualized storage play and where does what's your Flash strategy? Well our Flash strategy is pretty clear we talked about it in Project Hermes yesterday we are going to put Flash in the server and Flash in the server not in SSDs but directly on the PCI bus and the Flash in the server will operate at hundreds of thousands half a million IOPS it'll be extremely fast and with the Project Hermes we'll be able to make that content protected it'll be actually replicated to another site it has this software coherency across multiple sites so we're going to do that that's going to become a tier zero that'll be the active data of the highest performing systems in the world and customers will treat that like you're writing to a server memory and yet but it's protected by a device like Compellant so it's the merging of those two now we'll put Flash in other places we have Flash today inside a Compellant inside of Equalogic very specific use models that that plays in but I think you're going to see Flash become the predominant active data reservoir and when it does that being able to do it at the speeds we're talking about with projects like Hermes or Breakneck there's nothing faster we agree we've said for quite some time now unfortunately all active data is going to be in Flash and you're already selling the tier zero memory extension with partners like Fusion IO but it seems to me the key is the software that brings all that together that's right without the software without the software you can't get the reliability without the reliability no enterprise guy is going to trust it so we can get the speed without reliability that has improved to be a good model you need the speed and the reliability and software like Hermes is what does that okay and Hermes is the we're going to start seeing more of that I mean when does Hermes actually we're doing a demo of it we're doing demos of it this week downstairs and so our customers are getting to see it it'll be before this time next year excellent a lot of talking in the business Darren about big data you guys are not big data washing so you get a hundred points for that but at the same time there's some really cool stuff happening I mean you guys launched your VC arms I mean you have line of sight into some of the cool stuff that's going on in Silicon Valley and some here in Boston at MIT what's your take on big data is it really as big as everybody thinks it is and what's Dell's angle I actually think big data is a misnomer because we've been doing big data for a long time it's big analytics that is really what's interesting you have to have big knowledge about the data and so it's all about the software analytics again and so our currently we partner with companies like Hadoop to get those analytics but I think this is an early stage a lot of people talk about it but it's not about just having big databases it's about having the analytics to go with it you know we hired a general manager of the software business reporting directly to Michael John Swainson he's starting a software division he's already collected the software people within Dell that own the big data response and so I think he's got some actions and I want to let him introduce them but we'll really position Dell's big data of you very clearly here probably within the next six months you talk about kind of the enterprise functionality down to kind of commoditizing it for everyone and if we look at architectures for big data really talking scale-out architectures so we talk to Dario on the network and storage if for big data if you look at Hadoop and the Cloud Air and the Hortonworks guys they tried to get rid of the traditional storage stack so the question I have is if you look at architecturally do you think your scale-out architectures stack will be able to extend into that kind of space or do I need a lower tier more commoditized than what you have today? Yeah it's kind of bimodal today you have the guys going down DAZ very direct attached they can use any piece of storage and they offer all the software protection at a higher level and then you have the guys that really want the storage device to do this and if you start getting the tiering it's going to be logic and compelling so I think it's going to be a mix I think the quick response is with the Clouds going to happen and it's going to be big data it's going to be cheap and then you're going to have these very highly integrated very agile devices that are able to do business analytics which means you have more compute than just the cheapest machine have special software on there you have to be able to even that data has to be able to tier to the right location we're seeing the end of the classic array controller we're just seeing the end of the legacy one that is limited by its physical nature not its virtual nature Great answer along those lines what's Dell's play in the service provider market? Well Dell's got a big play in services we bought Perot and they had some excellent I'm sorry not service service like cloud service provider hosting software as a service the guys that are putting together if I remember I did the tour of the supernap last week in Las Vegas and I saw a lot of Dell solutions you have cloud services there but also the guys that are deploying software as a service are they using your storage products as part of their architecture? Yeah we're building several clouds at Dell and there will be one that's based on Aperture there will be one that is based on Equalogic because the Equalogic device can actually use that and there will be ones based on the tiering capability and there will be ones based on the super low cost cheap by the pound storage So you're getting the hook because you got a hard stop but we don't want to let you go so you know the red share we love the red share theme of course now you know there's no evidence there's no proof points that Ted actually hit it that far so to take the metaphor for forever here but so there actually is no video of Ted hitting that homerun I don't know if you knew that I didn't know that I just figured we're in Boston who's going to argue with the great Ted Williams? It's a legend so hopefully you can repeat that and have those proof points but we really like say we appreciate you spending some time with us we didn't get really much into the back up we love that vision think you're going to really drive some changes there congratulations on the progress you know like I said it's the second inning we appreciate you guys having us here and thanks for coming back in the Cube thank you for having me alright everybody that was Darren Thomas we'll be right back with the Cube live from Dell Storage Form 12