 Good day everybody, this is Dave Vellante, we're here at Wikibon headquarters in beautiful Malibu or Massachusetts and we're here to talk flash. Flash is taking the storage world by storm. EMC started it all in 2008 when it landed a haymaker by incorporating enterprise flash drives into its symmetric system at the high end. Now we all have been in this industry, those of us who have been in this industry for a long time remember back in the 1980s there were solid state disks, this is different. We're talking about persistent flash now coming into play at prices that are incredibly affordable and we're very excited to have a number of guests today. We're going to start off with Mark Sorensen. Mark runs EMC's flash business unit. He's the senior vice president and general manager of that division. Mark thanks very much for coming on theCUBE. Great to be here Dave, how are you? Good thank you. Again, we've known each other for a long time and you're moving up the stack as it were, right? You did a lot of work in the archive and the backup space and really got EMC's business going there and now you're on to a new project. Yeah, in February of 2011, Pat Gelsinger came to me and asked me to form a new business unit called the Flash Business Unit. At the time I had just wrapped up the VNX project. Before this, I ran the engineering organizations for Clarion and Solera, which soon became the VNX engineering organization. So I was ready for a new gig after that, that was very successful and that's doing extraordinarily well. And so when he came up to me and said, you know, I want you to start a Flash Business Unit, I was taking it back a little bit because even how close I was to the Flash Business because we had done so much software automation around Flash and we were selling so much Flash in our storage arrays, I wasn't quite sure, is this really something that warranted a business unit? And of course, as I educated us myself and what we're trying to do and we'll describe perhaps this morning, clearly it is an important business for EMC. Yeah, so take us inside of EMC because we had Pat Gelsinger on the cube in May of this year and he candidly said, you know, we fell behind some of the other guys that were out there. We started it all in 2008 and we shouldn't let that happen and presumably that's why he went to you and said, okay, we really need to go for it. So take us inside of EMC and share with us how this all came about. Yeah, and I think you have to really bifurcate the Flash business in terms of how it's deployed. Well, we're clear leaders in Flash in the storage arrays. We started in 2008 as you just described, deploying Flash in Symmetrics and then surely thereafter in our VNX platforms, you know, went through all of the shakeouts of getting the performance and the quality and of course cost, which has continually gone down dealing with that angle and became pretty successful. What really took off is when we began to introduce software that help automate the leverage and use of Flash by placing the hottest data on Flash devices called their data on spinning disk. And so this was the introduction of Fast, fully automated storage tearing and then subsequent and that was in 2009, in 2010 we introduced FastCache and FastCache is basically providing an extension to your DRAM cache with a lower cost Flash based cache in the VNX platform. And those two software products that really allowed you to most effectively and efficiently use Flash, so no longer manually trying to guess what's the hottest data, what's the working set that you want in Flash. It automatically did that by understanding your I.O. workloads, understanding what's hot, what's not and moving that to the appropriate tier of storage. So if you look at Flash in the storage arrays, I mean clearly we are the leaders. In 2011 we shipped more than 25 petabytes of Flash far and away more than anyone in the industry and if you think about Fast, in 2012 we estimate we have over 1.3 exabytes under management Flash. So tremendous success in the storage array. And I think what Pat was referring to where we kind of maybe not took our eye off in the ball but fell behind a little bit was Flash in the server, okay. You've seen some startups like Fusion I.O. really getting out in front of that and leveraging Flash in the server to really provide extreme performance leveraging Flash and so this was one of the reasons why we formed the Flash Business Unit and why we recently announced our first product from this Business Unit, VF cache. So what is the Flash Business Unit? Talk about that a little bit. What are the products in there and sort of the scope? So the Flash Business Unit has multiple responsibilities. Number one is really for the overarching strategy of EMC's Flash, Flash Technologies and Business. We made at the forefront of working with our partners that provide Flash products and technologies and as you well know there's a tremendous amount of Flash vendors out there. We produce Flash centric products, the first of which is VF cache. And I know after me Danny and Barry will talk in great detail of VF cache. In short, it is leveraging PCI based Flash as a cache in front of EMC storage arrays effectively adding another tier of storage, a caching tier to our fast strategy. So if you look at our fast strategy, you've got in the storage array, you've got Flash for high performance, you've got fiber channel for your warm medium data and then you've got near line satter devices for cold data, great economics, great performance. We've now augmented and extended that by leveraging Flash in the server. So now you have an ultra high performance tier, a caching tier where you get the best performance that you could possibly get using Flash close to the CPU on the PCI bus, but all of the protection and intelligence that an EMC back end storage array with fast and other capabilities offer. So that's our first product. We are building other products as well, nothing that we're willing to share too publicly. I'll give you some hints. Yeah, show us a little leg here, come on. So a project which, well, you know, the codename for VF cache was lightning and as you well know, what follows lightning is thunder. We all know that if you've gone to, if you're a meteorologist fan or what have you. And so we do have a project called Thunder and it is, as you could call it to some degree the big brother of lightning, it's more shareable, it's more scalable. It deals with not only reads, but writes very high end, high performance for those applications that require the utmost in terms of low latency. And so that's kind of, again, you know, thunder follows lightning and we get a few other things up our sleeves. Mark, it was interesting to hear you when you were describing the initiative around, you know, VF and what we would call project lightning before the announcement. You talk a lot about software. Now, so your business unit develops new products as well as you have to leverage existing products like Fast within EMC, is that correct? It's correct. Okay. And so are you guys a P&L? We do have a profit and loss statement, that's right. We are measured as a business in terms of expense and revenue. Yeah, so the interesting thing about the strategy, if I understand it with your approach is you're, it's not a rip and replace. You're really trying to extend the IO stack, if you will, toward the server. You talked about that earlier. What happens to the existing storage infrastructure? What are you telling your customers, the Symmetrix customers, the VNX customers? If you look at those customers, right, certainly performance is a critical attribute of the need for storage and information management. Just as critical and to some degree more critical is availability, scalability, the reliability and disaster recoverability of your information. If you ask any corporation, what's their two most important assets? It's their information and it's their people. And so while performance is very important, performance means nothing if you don't have those other itties, if you will. Obviously, our customers want us to take their existing investments in VMAX and VNX and Icelon and to leverage new technologies to extend and integrate it into those highly reliable, well-proven environments. Environments that they've got a lot of investment in, whether it's training their storage administrators and managing it or customizing their applications to take advantage and leverage it, etc. So it's not a rip and replace, you know, it's an augmentation and extension of the existing environments and really leveraging the best of both the new technology as well as the dependability of our existing platforms, which frankly our customers have invested a lot of money in and it's been proven over many, many years. So do you see Flash becoming the home for all active data over time? Well, so I think there's a couple of things here, right? I think you will see Flash leveraged predominantly for any data that requires some sort of high performance. Now, where you cut that line in terms of is it active data? Is it hot data? To some degree, what we do is give customers the choice, okay? By using leveraging fast technology, you can kind of decide how much high performance infrastructure you want to place on for your environments. And then we know that if you, and we know this because we live this every day, is that, you know, data, not all data is created equal. And not all data stays in terms of its importance and its use static over time, over time things age, they go out of, they become cold yet in many cases you need to keep those and retain that data for years and years and years. And so our strategy is really is to write infrastructure at both cost and performance. So we're not extremists, right? Some extremists will say all Flash for every single thing. And so that, well, it sounds interesting and at the end of the day, if you look at the economics and you really have to really look very carefully of the economics because there's lots of folks out there playing pretty fast and loose with the facts and they'll make claims that their new fangled storage arrays can be just as economically viable as spinning desk storage arrays. But when you look back on the covers, it requires extraordinary results and deduplication, meaning the average ratios of deduplication are not going to get you there. It requires lots and lots of funny things, unproven things on how they deal with RAID. They ignore the replacement factors, particularly if they're claiming they're going to use consumer MLC of pulling out discs and sticking back in as if it's almost a consumable, putting aside whether our customers actually want that and want that type of data center there. So there's a lot of fast and looseness with the facts. I think at the end of the day, spinning discs don't go away. Certainly not in yours and my lifetime. You just look at the numbers, flash costs, while they have come down tremendously. Since we introduced them in 2008, you never see a crossover point with hard discs. Hard discs doing a great job with density, right? And of course we see flash getting less expensive. We're seeing smaller lithographies but of course with those smaller lithographies comes issues. Issues with endurance and indeed issues with performance. So I think again, you know, we try to take a balanced approach, not an extremist approach that some of the startups are taking. And I think our customers who have depended on EMC in this logical, straightforward manner, really are embracing it. So you're basically saying that the customer decides, defines essentially what active data is and what goes on flash as a function of both performance and cost requirements, right? And the role of the sand maybe changes, maybe not, depending on what that economic and performance value equation is. I think you're seeing, excuse me, I think you're seeing data move closer to servers and servers moving closer to data. And so as we drop a PCIe flash card into a server controlling it through fast and other capabilities acting as that read cache in front of our storage arrays, we're getting closer to the server. And one of the reasons why we do that is because we get better performance, right? If I'm sitting on a flash, on a PCIe bus with my flash card and I'm serving up reads from there, I'm going to get the best performance I possibly could. It's kind of a physics thing, right? Yet on the back end, I mean, all the investment customers have with their SRDF deployments for disaster recovery and their data protection scenarios with data domain or Avama, all of that stuff is extraordinarily important. You just can't throw it out, right? It's just crazy to do so. So we try to give kind of a balanced approach and a more systemic system view versus just kind of a one-trick pony approach. So that leads me to my last question for you. Is EMC a systems company or a storage products company? I think we are a solutions company and we provide storage and information solutions across an increasingly broad spectrum of solving more and more problems and some of this obviously is a natural evolution and I think flash is really a natural evolution. Others are going into more adjacent markets like big data, which we've found to be very compelling and very interesting. So I would call us a solutions company. Certainly, our portfolio has got an extraordinarily broad, we're solving more and more problems that we haven't in our past and I think you'll see more of that over the coming years and months. Yeah, we didn't get into the big data thing. I want to do that with Dan and Barry, but I mean, you're really seeing new types of applications emerge and flash is enabling a lot of things. Absolutely, absolutely. And flash is certainly, when you think about scale out web applications, but also your bread and butter oracle databases, SQL databases, it's really dealing with both and flash certainly helps with the scale out applications that we're seeing in the 2.0 web environment. Exciting times, follow Mark Zorinzen, right? You're in the heart of the data protection business and that took off at EMC, the whole mid-range unified business taken off. You're now at the epicenter of what I think is the most exciting trend in storage and that's that intersection of software and flash-based architectures. And so congratulations on getting this off the ground and your new role at EMC. And thanks for coming inside the queue. Appreciate it. Thanks, everybody. Thanks for watching.