 Lazz, thanks for taking the time. I'm John MacArthur. This is Callie Lewis. Hi, nice to meet you. I appreciate you coming in. I think this is your first time on the queue? Absolutely. Okay, all right. Oh, me nervous? Not really. Okay, good. Well, you could make me nervous by asking the wrong things, but okay. No, actually I'm going to be very careful and ask you not to disclose anything that you shouldn't, so we can have you back again. But I do understand that you helped lay some concerns among customers that regarding an investment roadmap on the Equalogic platform and things like that. Can you just talk a little bit about that? First, can you start with what were some of the concerns? Well, so... Actually, you know what? Start one place first. Tell everybody who you are. Okay, sorry. I know we were just so excited to have them on and we just went straight into it. Because I've known and followed you for a long time. You don't know me, but I know you. It's one of those things. So let me give you a little bit of background. I actually was one of the four or five architects that actually built the original Equalogic product. And I joined Equalogic in early 2002. Right. I'm mostly known for the guy. They keep calling me the gooey guy, but I run the software development teams that own all of the control plane. Everything that controls the security, the automation of the features in the array that everyone's known to grow and love. And then pretty much all the software development that goes on in the various host integration activities. So anything that goes on Windows, in VMware, ZX, on Linux as well. And pretty much any other ecosystem-related software development. So it's actually a relatively large group now. When I started, I was me and two guys. And now we're about a hundred people or so. Can you hold a few patents here? Yes. We've discovered a few interesting ways of doing things over the last ten years. And, you know, things around snapshot technology. Our multipathing is also another area where we have some intellectual property that's very unique. And there's a few other things ongoing right now. And actually the interesting things about patents is that they're public, so I can talk about them. We have some automation capability that we think is unique around auto-logins of ISCSI initiators. And there's a number of other areas that I can probably do numerous to mention. I'll bore you with it. But anyway, a lot of stuff there. But relative to Dell having a lot of IP now, you're bringing in a fair bit of IP from the ecologic side is the point. Yes. And this, by the way, is Dave Vellante. Nice to meet you. Nice to meet you. Pleasure to meet you. So are you from New Hampshire? I actually live in Boston. Oh, you do? I grew up in Boston. How about those Bruins? That was awesome last night, wasn't it? Absolutely. It's a series again. As Cedric Maxwell would say, that's the big needle in the arm. We're back in this thing. Well, growing up, I just wasn't... This season would be over by now. So hockey in June is still something that I haven't mentally really grasped. Well, you're younger than I am. In 1972, I remember, because it wasn't June back then. It was a little earlier, but they stretch out the season. So what did you do? You commuted to New Hampshire every day? Exactly. Exactly. So take us back to that. I mean, how did this collection of people end up in New Hampshire in this renaissance of storage? We were talking to Phil Soren earlier about you had companies like Equalogic and Left Hand and Compellent and 3PAR and Data Domain. It was like this storage hotbed coming out of the dot-com bust. Sure. And you understand Silicon Valley, right? That makes sense. But how did it happen? Minnesota, right? New Hampshire, how'd that happen? Well, so there's a couple of interesting technology areas in New England. And I came from the networking industry. As it turns out, if you recall in the late 90s, there were a lot of networking companies that came out of the Route 495. Hey, I'm Banyan Systems. I worked at Banyan. You're Banyan. That was actually my, I thought it would be my favorite job ever. Yeah, something happened. Well, yeah, Microsoft happened actually. And, you know, lots of storied companies that were networking related there. And the dot-com bust, people forget, you know, people keep thinking of pets.com. The networking vendors were the real money makers during the late 90s. Yeah, there was a bandwidth craze. Exactly. And that was the area that was substantially overbuilt. So when the boom busted, there were a lot of us networking geeks out of work. You know, what's ironic about that is the story, and you remember this back then, was that during the gold rush, the people who made all the money, it wasn't the people who mined the gold. It was the people that provided the infrastructure. Exactly. And so that's why the networking business exploded. It was like, oh, it's going to be the same thing. Exactly. Well, so I was at a networking startup and it went out of business. As it turned out, maybe two months before I joined Equalogic. And, you know, I got a cold call from Paula Long at that point. Well, you did? Okay. And she was looking for people that understood, you know, the control planes and system management. Did you know Paula before? I had never met her before in my life. And famously, you could check with my wife on this. I had told her after Innovate Networks closed that I would never join a startup again. I never wanted to be a people manager again. Never say never. And then, you know, two months later, I'm showing up at Equalogic working for Paula. And Paula had, you know, this interesting way of, you know, collecting people. She would spin you. She'd listen to what you were saying. And somehow you would leave the room thinking you're going to go to Equalogic and it was your idea to begin with. And it was very, very, very strange. And so I learned a lot about management from that approach. And so, you know, she actually took a few people from Networking Space. Okay. There are obviously some relatively large, actually very large storage companies in New England. I've heard of some, you know. Strangely enough, we did not really hire a lot of people that had storage backgrounds. There were some people... That's why it's a product so simple. We have... Clearly the lower level, you know, raid code and the storage stack did come from a very experienced, actually it was people that worked at Adaptek. Right. No, there was a big Adaptek community in Nashville. Exactly. Prior to the start of Equalogic. It turns out that we got the building Adaptek also. I actually visited it when it was Adaptek. I didn't know that, John. Oh, yeah, yeah, yeah. Adaptek had a whole... Is it a great... Grant who set that up? ...that had the raid controller development was done up there. Exactly. And the very same people. So they actually closed that office and leased that building to Equalogic at some ridiculous per square foot rate, including all the furniture. Right. And so there's a select few people that went to work for a new company in their old cubicle. Very nice. And so it was a very interesting... If Adaptek had been really smart, they would have leased you the building and the people and never laid them off. And then they could have had a piece of a nice accent as opposed to being sold for pennies on the dollar. Yeah. So moving on from that, we actually always had this focus on simplicity. And because none of us were jaundiced from having worked at storage companies, we were always looking at these workflows that, you know, configuring a sand. You know, all the crazy things you needed to do in a fiber channel sand at the time. It would drive us nuts. We'd have to sit there and scratch our heads and say, well, why is it this way? And, you know, Iskazi is simple and we were all networking people. And so a lot of the provisioning model that you have in the arrays today came out of those discussions. What about on the monitoring side too? You've got provisioning and you've got monitoring. And monitoring is a big part of networks, right? Exactly. And you'll notice that our array from the very beginning, and this is my fault obviously, I have put an SNMP agent in the box. So the array is actually monitored like a network device because it's an IP device. And even today, you have a really wonderful tool that we give away for free to our customers, Sand HQ. It's an SNMP monitoring tool. And it's the kind of thing that if you had a switch or a router, you'd be used to having. But, you know, for some reason in the storage world, this is not there. You know, you have all sorts of arcane ways of getting this type of data. So, but coming from a networking side, were there any surprises for you when you got into storage? It's like, gosh, I wish I had known that. You know, or was it really, or did the storage world actually over complicate everything and it really is that easy? No, no, no. I don't think that they over complicated things at all. I think that no one had, you know, people build up sort of a frame of mind or a frame of reference about how they're going to use a product. And, you know, putting together a raid set and, you know, and deciding what kind of a raid type you're going to have. There's a whole bunch of workflows and then they make a lot of sense when you have a very, very specific application in mind. And you had this big long decision tree and it seemed over complicated. You know, when we finally shipped the product, I started to realize as I got feedback from customers that, well, there is a reason they do that. They want a raid 50 or a raid 5. They're looking for certain performance characteristics, certain kinds of applications, things like that. Exactly. And that was sort of the point at which we stepped back and said, well, wait, you know, maybe we do need a few more knobs. But, you know, overall, the philosophy always was that we're going to try and make as many of these decisions for our customers as possible in the code. So Darren Thomas often calls us the apple of storage arrays because the first thing we did is we took all the knobs out. Do you ship without a user manual? No, we do ship with a user manual. OK. So you're not quite to Apple yet. How thick is it? Is it on 11 by 17? How much is it weigh? Well, the litmus test, I'll tell you the litmus test always was that no one who actually tried to test the product, we beta test with our user manual. No one who actually sees the product for the first time, any release has a user manual, and we watch them closely. OK. And if they get confused, that's where you have to write. Well, no, actually, it's caused for alarm, and usually it becomes my problem. So that's where you have to code. Exactly. So we write code to make it simpler, make things more obvious or automate steps. OK. And that's always been the philosophy for us. I remember when I was at university, a lot of people put down sidewalks based upon what seems to make sense to them. The university I went to put down the sidewalks after the students beat the paths to the doors. They saved a lot of concrete. I think that's a good analogy. Students are going, where's the bloody sidewalks? They'll show up, just wait to see where they're going to walk. Find the pattern. So when Dal was doing all this back and forth with HP and 3par, what were you guys thinking? You just sort of sit back and wow. You know, we were amused. We knew the 3par guys. And actually, Mr. Farley. You would know Mark, wouldn't you? Yeah. So Mark was actually our blogger at Equalogic. At the time of the Dell acquisition of Equalogic. Yes. And then he moved on to 3par. And I thought it was the funniest thing. So I emailed him and I said, welcome back. Turns out that was premature. So you guys were kind of like, in that same, you know, Equalogic, left-hand, compelling, 3par, sort of in that same class, if you will, Superstars. But taking different majors. Did you watch each other? Or were you more like, we know what the target is. It's that big gorilla in Hopkinson. It really was. And if you look at the storage market, there's just so many different layers of submarkets. And we were always focused on the lower to mid section of that market. And 3par was, you know, from mid to high. And, you know, the compelling also was actually more overlap between us and compelling. But they were more or less in the same ballpark as us. But, you know, we were always watching each other and we were always looking at, you know, what feature does 3par have and what feature does compelling have. And if you look at the feature roadmaps for all these companies who introduced what, you'll see Equalogic introduce a whole bunch of things first. But immediately after the introduction of an Equalogic feature, compelling would announce. And then, you know, vice versa. And so, you know, there was definitely kind of a sibling-friendly rivalry. It's amazing, isn't it? That innovation occurs that way. You're on, you know, different parts of the country. In case of 3par, different coasts. Sure. And yet you come up with similar concepts, applied differently, maybe applied to different markets. But, you know, it's what you see in our industry. Oh, yeah, absolutely. And if somebody thinks of it, somebody else is thinking of it at the exact same time and maybe thinking about applying it differently. Well, what happens frequently is that, you know, I think it's those sales guys that do this, right? They go around and they talk about a particular product to a customer and maybe the 3par product was not all that great for customer A, but, gee, that then provisioning stuff seems really good. Right, right. And then, you know, Equalogic comes in afterwards and, you know, they start yelling at us because we don't have thin provisioning. And then, guess what? So that's how the cross-pollination. So you're watching 3par and there was a big gap and then that fell apart and then all of a sudden the compelling comes into the mix. As you said, there was much more overlap there. Was it like, okay, all right, well, wait a minute, you know, how's that working internally and what are you guys, you know, how are you, are you even worried about it? Are you trying to reconcile that or is it more just like, hey, let's just sort of see what happens in the marketplace or are you being more deliberate, I guess? Well, the first thing, we've met with the compelling guys from very early stages of that deal. I think, you know, we are very friendly with them. We actually like them a lot. They're very smart guys. And in some areas, they solved problems very similarly to Equalogic. In other areas, you know, we're vastly different, but we're not really that worried about it. We're spending a lot of time thinking about ways in which we can leverage each other's work. And one of the sort of truisms of integrating companies is that it takes a long time. And they've been building stuff around their system for, you know, 10 years almost, and we have a long history as well. And you can't simply pull parts of systems together by ripping code and compiling it again. It actually has to be well thought out. I mean, sometimes I wonder about the ROI of integration. I mean, everybody wants integration, but except it's hard. It's hard. I mean, sometimes it's better maybe just to leave things alone. What are your thoughts on that? That's actually a very, very good question. And I have, you know, I'm very opinionated on that. I think integration for integration's sake doesn't necessarily solve any problems. It creates more problems for the vendors, right? You know, and at some point, we're competing in different markets. And the requirements of one set of customers should dictate what that product, what the product they use looks like. So having uniformity and complete integration is not always desirable. And it's hard. So you end up, you know, in some cases you could easily end up building something that's very hard to build and not really derive much benefit. So you have to be very surgical about where you integrate. So what are the big differences? Is it channel? Is it use case? Is it capability? I mean, because you're both simple product. You talk to both customer base. It's like, oh, it's so simple. I get one equal logic customer this week said I got my weekends back, you know, and other guys at Mazda were saying, you know, how simple, you know, a compelling made their lives. So you see a lot of the same themes. So there's some similarities there. What are the differences? Well, you know, you know, compelling obviously is a much bigger system in terms of in terms of scale, you know, scale, right? We are more focused in storage as small increments. So what you'll find throughout our history is that our key customers were the guys that were buying their first sand. They had, you know, seen the light and they really realized that, you know, the shared storage thing is something that we need to do. Probably because of VMware, more than anything else. VMware, as you know, has kind of driven the whole sand market over the last few years. And we have always been there as the, you know, the most expedient and lowest cost and easy to use solution for that constituency of customers. And compelling, and this is, you know, I don't have hard data on this, but compelling seems to have much more traction in more established jobs where they've had, they've had shared storage for a while. So difference in approaches from a market standpoint. So they really are different. You wouldn't buy a compelling if it was, you know, your first experience with sands more or less. I mean, I'm sure they can, you know, specify a system. But that's the way it's been positioned for the most part. And, you know, we really didn't see a lot of compelling out in the field when we were pitching. You didn't compete against them? Yeah, from time to time you'd always compete against them. The big boys were the ones we were constantly competing against. Right, sure. And one of the things that I heard during the conferences, if there's no overlap, there's gaps. So it's true. It's okay to have a little bit of overlap. Sure, sure. It's not the only company that's experiencing a little bit of that. Sure. You see, you even see some of that in EMC. They try to manage it through marketing and through sales and through channel and things like that. But, you know, there are workloads that Echlerion does, that Asymmetrics Cadours or vice versa. Exactly. And you have to have an informed, you know, process for how to select a product, especially if you have multiple choices. It's all about the customer. And so, you know, I think as we go forward, you know, the main points of integration between the products are probably going to be in the area of migration. Because, you know, the data migration across systems or within systems or across systems probably as a start. It would be very nice. I know a lot of customers have approached me this week saying it would be very nice to be able to use one or another system to hold replication data. There's a target. We're going to replicate from one to the other hand. And, you know, that, fair enough, that's actually a good use case. And we're definitely going to be looking at things like that. So for example, I might not want to have a target, a remote target be the same. Maybe I got a, maybe I got a compelling in my data center. I might want to have a remote target that's an equal logic. Exactly. A simpler, lower cost maybe. Exactly. I may have some really smart people in my data center, but maybe I don't have IT staff in a remote office. I need something that I can just sort of leave there with the lights off in a closet. You know, that might be possible. So Lazarus, you came from a networking background. How do you like the storage business? You know, it's, well, the networking, the networking business is very exciting in the 90s. And, you know, that all seems to have gone away and now that, you know, it's quiesced to be really one large company. And the storage business is actually as exciting as the networking business was in the 90s. Because you don't have a, I mean, EMC is the largest vendor. I think it's got, I don't know, 30% share or something. Why do you think it was that Cisco was able to capture whatever, 70% share of the marketplace? And that never happened in storage. What's your observation there for somebody who's been in both worlds? Well, the first thing that we should point out was that Cisco picked up a lot of their market share as a result of that .com bust. I remember a lot of the infrastructure vendors were very, very, very extended at the time. And you'll recall who were the competitors. Nortel, where are they now? I mean, you never actually had that happen in the storage industry where all the competitors just folded and were in such bad shape that they weren't even able to compete. Yeah, maybe almost. I mean, they unfolded, but EMC and NetApp came out of the .com. They were in trouble, but they rebounded. They were in a little bit better shape. They were in better shape. And so they continued to hold on. You had a lot of, I mentioned Nortel. Do you remember Cable Tron Systems? Yeah, absolutely. There were a lot of 3Com. 3Com kind of disintegrated. I mean, 3Com was going to get into the storage networking business there for a while. I remember a couple of guys that I used to interact with that were trying to grow the storage business at 3Com. It didn't really take off, but you know. They're in the storage business now. Yeah, that's right. It's true. And of course there's lots of rumors around networking here at Dell. So I won't, you know, how far we're going to go. I will. The answer's going to be. No, I don't. We wrote about that. We're on record of saying this. It makes no sense. And you're in the networking business today. Dell has. Oh, yes. And actually, you know, this is one of the more interesting things that I think gets lost on people. You know, Ice Guzzi is probably the biggest driver for high performance networking here out there right now. I mean, most networking vendors really love applications that demand a lot of networks because it requires you to buy the best gear. Right. Ice Guzzi is one of those applications. You're definitely driving some network demands here. I'm going to slide out and let you continue for a little bit. I've got to run to another quick appointment as it was great to spend some time with you. Same here. All right. Cool. So let me just tell you out there we're live here at the Dell Storage Forum. This is day three for us. This is SiliconANGLE and siliconangle.tv's continuous coverage of Dell Storage Forum. And SiliconANGLE is the worldwide leader in tech coverage. This is theCUBE. How do you like theCUBE, Lazarus? Well, it's my first time in theCUBE. I'm glad you invited me. This is fun. It is fun. theCUBE is a place where we try to find the smartest nodes, extract knowledge, and learn and share with our audience. And so we appreciate you coming on here. What's the future hold? If you look out and put on your telescope, what do you see happening in storage? You saw, let's take people through. You saw the old mainframe days seated to the open system big box monolithic days. Then sand came in. And sand kind of, it did some great things, but it really didn't deliver on the promise of simplifying storage. It definitely helped with disaster recovery. It helped with sharing. But it certainly didn't help with simplifying. That's where you guys came in. That's right. And some of your peers. What's next? What's the next wave in storage do you think? Well, there are a lot of interesting developments. So first of all, the big driver in storage right now is virtualization. So anything related to enabling better virtualization is interesting fodder. So you mean server virtualization or storage virtualization? So as I mentioned before, server virtualization, VMware, Hyper-V, those types of solutions are actually drivers for sand storage. The way you make those things work, the way you achieve the level of availability that you desire is through shared storage. And it's causing our customers to go out and implement things that 10 years ago they wouldn't have considered to implement clusters. For example, you've been around a while too. Clusters in the 90s were this crazy thing. Supercomputers. Exactly. And now it's relatively commonplace. It's taken a while, but if you look at when the pickup occurred, it was virtualization. Now going forward, in the virtualization area, you have a whole bunch of other performance bottlenecks. And in order to alleviate those, you're going to see the next generation of innovation. One of the big things is SSD and flush. It is kind of a game changer if you think about it. We've built a huge amount of infrastructure around managing, rotating rust. And at this point, something like that technology really makes the infrastructure well, kind of creaky. What are we going to do about having something that behaves almost like memory? Everything, all the code that we wrote needs to be rethought. And there's a whole new set of paradigms. We're not going to use flash like we do disks. And so there are all sorts of use cases for flash in areas, especially in virtualization, where you have really hot data or data sets that are incredibly concentrated in some way. We actually have an array right now in the equilogic array, the XBS array. That's half SSD, half rotating media. And that is actually a wonderful product for virtualization. But there's a lot more we could be doing there. In terms of optimizing the flow and performance and cost. Exactly. So you mentioned several. Virtualization, clustering, SSD and flash, this notion of hybrid arrays. Can we throw in desktop virtualization as well? Is that something you've seen before? That's a very, very big area. And the key driver for that is economics. And as an industry, we're going to be challenged to come up with ways to make the economics of desktop virtualization work. When you look at the cost of desktop, it's actually pretty cheap. Cheap hardware. There's real management cost in managing that stuff. And by centralizing it, there's some leverage to be gained. But you still have to obviate all the costs that are inherent in having all this big infrastructure to manage desktop virtualization. So a lot of these technologies that we were talking about help, we want to get to the point where a virtual desktop costs less than a physical one. And we're getting there. I've been talking this week, we were at Citrix Synergy a couple weeks ago and what struck me there is that the whole notion of desktop virtualization, even though everybody uses that term, seems to be outdated. It's a very kind of 20th century concept and now with mobile computing and the age of smartphones and iPads. That's maybe it's the mobile user, the mobile applications, the mobile data that maybe changes, maybe shifts the discussion from one of economics which really don't hold up except in certain use cases for things like VDI. Maybe shifts that discussion from TCO maybe to value. When we first got PCs we weren't saying what's the economic justification for PCs? I was like I've got to have a PC because it's going to increase productivity. And so maybe I guess we should call it client virtualization. I sort of got thrown off by the word desktop there. Yeah, client virtualization or even application virtualization is really what it comes down to and accessing your data from that mobile device and being able to manipulate it. That is another area, a big area of innovation these days is mobile arena what the ultimate form factors of those devices are going to be and how they're going to use storage. That hasn't been decided yet. You're starting to see a lot of people around the conference are carrying iPads so obviously there's a need for something that looks like a notebook that's a touch screen but the applications for that and the really big data centric applications for that they're starting to come online. How are those things going to connect to data? Is it going to be over the web? Is it going to be something more direct? There's some interesting questions to be asked there. And how about another area a great perspective so I'm just going to keep going. Another area I'm interested in is data protection. I've been on record as saying backup is broken and I really think that data protection as a service is the fix and what I mean by that is really an outdated model of expensive insurance. I got one size fits all, I do daily incrementals, weekly fulls there it is, take it or leave it and I'd like as a business person to have more granularity in the backup choices that I have. I'd like to pay more for my mission critical applications and not pay as much for less mission critical. And of course everything's moving to disk base which brings us to you guys. Do you see the potential to play a role in that transformation along with virtualization and cloud computing? Oh absolutely and that is an area that we think is very fertile for better solutions. And again, going back to the previous discussion we had about how we changed the way the workflows of storage were happening by developing ecological. There's a very similar set of sometimes bizarre arcane workflows in backup. Why do people do that that way? And it all comes it centers around habits and old systems and tape. A lot of process that's built up. Well and if you think about it even if you take away the tape and put in disks if the process is the same have you really done anything That's what Data Domain did and they succeeded wildly doing that Data Domains that hold the tongue-in-cheek motto was tape sucks Well it turns out that backup sucks That's really exactly There's an opportunity there I think you're right, I think it is fertile ground. And so there are lots of things when we first built the auto snapshot manager to do VSS enabled snapshots of exchange and SQL server we had a very non-backup use case in mind whenever you're sitting there and you go away at your application you do something really stupid and you break it and you wish it was five minutes ago We wanted to be able to dial it back really quickly to five minutes ago and you can be on your way it was like a giant magic eraser that was the use case but as we developed it and showed it to customers everyone kept asking us wait a second these snapshots they're backups aren't they can I throw away my backup software and we've always been like we can't, right? Because backup is one thing and recovery is everything Exactly, exactly What was interesting about those snapshots is that we automated the recovery from them so it's quick and easy so they saw this and they said well that's really what I want for my backup solution that should tell us something because it's not that the customer has an idea really that is particularly incisive but they're telling you their problem in an odd way in that case and we've constantly been looking at that and saying well gee, is there a set of tools that we can build around backup to make that whole process more intuitive and more useful and if you think about it there are all sorts of different models but ultimately backup it's not just getting back your data from five minutes ago it's also two years ago Even though you don't do that a lot you don't have the ability to do that just in case Exactly, and so if we sat down and we thought through a process of what backup should be I bet you we'd find that we drew out a process that's completely radically different from the way people are doing it today All right, Lazarus of Vechiarides am I saying that probably? Vechiarides, thank you very much for coming on theCUBE and sharing with us all your knowledge and it's great to have you good luck with everything I'd love to be back. All right, excellent.