 Hi everybody, we're back. This is Dave Vellante at Wikibon.org. I'm here with David Floyer, who's my co-host of this segment. This is the spotlight where we're doing a deep dive on software-defined storage. We've heard a lot about that this week and this year, actually. VMware Last VMworld in Las Vegas in August of 2012 really introduced the concept of the software-defined data center and software-defined storage really started to come out. David Floyer, we've talked about software-led storage, software-led infrastructure. I guess that term didn't stick, but it sure does within our community. So that's kind of our generic name. Dale Deegan is here. He really is focused on software-defined storage within HP's storage division. Dale, thanks for coming on theCUBE. Thank you very much, Dave. So you just heard Craig, he's kind of set up at a high level software-defined storage, what it is. I like this definition, it was clean. David, I think it's pretty consistent with our definition. Absolutely. Sometimes some of these definitions get a little too convoluted, so would you add anything to what Craig said or any color to sort of your view of SDS, Dale? Really, Craig covered it well. I think it needs to be open enough that it can be inclusive for the entire industry on that. And really just covering it on saying that we're taking industry standard type solutions and you can take the data services and apply those to servers or storage and then also orchestrate that management. I think that's a real open type of discussion that allows for lots of room. This event is pretty exciting because you've seen lots of different people talk about software-defined storage and I think we're in a good spot. We've been doing this for a lot of years. I feel like we've got a real proven track record with applying this in the industry before that term even came about. Well, it's kind of interesting because a lot of times companies will, startups will innovate and they'll be working their tail. And all of a sudden the market just comes their way and the wind is at their back. So you were at left-hand. You probably didn't use the term software-defined storage when you were at left-hand. It's a very prescribed term. So what, I mean, take us back to some of the thinking in the early days and did you see it in parallel with virtualization and the trends there and did you see it evolving into something like this? What did you see at the time and contrast that with where we are today? Well, I think the original founders of left-hand networks were looking at storage, and this was 10 years ago so they were looking at all the enterprise data features available and basically saying how can we bring this price point down to where more people could have access to this, to really bring it into a nice price range and with that they focused on a concept that I think just fell into the right spot at the right time and what I mean by that is they brought data services to commodity hardware that they could build and really just port from one X86 platform to another. The timing of that dovetailed nicely with the growth of virtualization. As the server virtualization grew, so did the growth of left-hand networks and about five years ago we were acquired by Hewlett Packard and really it was a wonderful fit. Hewlett Packard has the best technology for servers on racks and blades and at that moment in time I saw that this is destiny. There's some pretty exciting things happening and it felt like I was the only one beating that drum for a little while on that and to be honest and it was fun in the last few years we've gotten a lot more where people understood it and again timing is everything. About a year and a half ago we made a decision to rebrand this technology as Store Virtual VSA to be part of our HP Storage portfolio and really try to re-educate the world about the possibilities of being able to port this software not only to X86 but also as virtual machines into different areas and so we actually planned that we were going to launch Store Virtual VSA and talk to the world about real case studies of real customers that have been shipping it for a long time at VMworld and like I said timing is everything. Last year the theme was announced of software-defined storage and so now just fortuitously this has become a real rally point for us. We are able to showcase our leadership and the proven technology of what we've done with not only Store Virtual VSA but also we're expanding it across the portfolio to all of HP Storage and actually we have a new release of Left Hand Operating System which we're showcasing on the HP Store Virtual VSA. We announced this last week with a few new features. So I'm curious about why you went to virtualization because if you look at the history of virtualization storage was always the bottleneck because of the blender effect and all of those. So why did you choose to go that route rather than sticking on the hardware itself? What was leading you to that direction and what did you have to do to make it? Sure. I think there's a number of factors on that. As I spoke of the historical legacy of Left Hand Networks applied a Linux-based operating system onto XA6 servers. So now what we're able to do is we're able to take that same operating system on a physical array but also add it into a virtual machine and we've done that since 2007 and at the time there's an anecdotal story. It was created because the developers didn't have enough resources to test the code that they wanted to work on. So they said, oh, I'm just going to make a virtual machine and I'll start, I'll run a bunch of these because they didn't have enough servers. And we walked along and one of the product managers saw that and said, what is this? And wheels started turning and since we've been shipping since 2007, originally I think it kind of got sort of a lower class ranking because the servers and the technology really, it was kind of ahead of the game on what was the underlying hardware. It would consume possibly the entire server and so then there's no value in virtualizing. But as Craig had spoken to earlier, as we see the evolution in HP, we have the Gen 8 based servers and other companies, there's other servers, the cores are growing. The efficiency of the hypervisors is much, much more. The price of solid state is shooting down every day. The size of the actual drives is going up and so the economics of that is all coming together to where now we've have, it's quite reasonable and it's quite smart to take and have fault tolerant applications, VMs alongside their fault tolerant storage and the scale out architecture of the store virtual VSA allows you to linearly grow performance and capacity and it's just perfect fit with virtualization on here. Originally it was the perfect scale out architecture alongside with the physical servers that were virtualizing. Now we can combine it all together. You're seeing that in a number of the solutions here on the floor and I think HP's really in a great position here because we've got the best servers and we can align these forces to where all of a sudden now our customers can build their own converge systems within a few servers. They've got their servers, storage, backup, all of this taken care of in one solution. So part of the software led or software defines a storage or anything else is that you yourself can be managed by an orchestration layer or whatever layer is above that. Can you talk a little bit about the strategy that you have in that area? What are you providing for your customers to? Enable that? Yeah, I think the definition of software defined storage we spoke of earlier really splits into an orchestration aspect of this storage and also the data services and I think HP's nuance is that we can package those data services as a software package that can be applied to anything. Now the orchestration aspect of that because the left hand operating system was ice-scuzzy based really we've been involved with the open stack over the past few years since its beginning and HP's continued to have new and increased funding and development into that and so moving forward, open stack is a vital portion of our strategy because we want to give people options on how they can orchestrate and build their own clouds with this type of solution but in addition we're building an overarching management structure for the storage portfolio that's in parallel. So every effort we do in open stack will also be available at a portfolio level and as we look at this, I've been very fortunate position. I've just happened to be lucky to be in left hand networks at the moment in time that we were acquired. We're lucky to be acquired by HP, the best server company in the world and now we're looking forward and saying we can use all of this experience, all of these lessons learned and expand that across the converged storage portfolio. So not only can we virtualize in software defined storage primary storage but as Craig was speaking of, we have StoreOnce that we just recently released for DDoop and Backup and you can see a time where today we package it as VSA but I don't think it always has to be that way. I think we can have a portion of software that can be applied to servers and can start off with primary storage, give you high availability with lots of data services to be able to give you disaster recovery for those servers and then maybe within the application you can actually upgrade the features, you can grow your capacity, you could grow a feature set to have DDoop and backup applications and maybe even block and file and I believe that's where I envision us going. We're going to have one piece of software that will be unified across the storage portfolio. It will be the end of the spectrum. We spoke of earlier how you've got three part with amazing ASIC technology and that's one end of the spectrum that's focused on service level agreements and guaranteeing and always will guarantee the top performance. The other end in the cost optimized area, I think we can have a shared data services from one end to the other to where we have a software solution that covers all of those areas and at an orchestration layer there's unified management among all of those. It looks at a physical array and a virtual array as peers and additionally has inputs not only into VMware and Microsoft with the application integration but also an orchestration through such things as OpenStack. So can you unpack the announcement you guys just made? Sure, so last week we announced that the left hand operating system 11.0 was released and we did something new. We launched this specifically on the store virtual VSA as the first wave of this and because of, we did that because of some really interesting technologies. We've added in adaptive optimization which is sublun tiering on a per node basis and so now you can go out in a virtualized environment and not only provision any servers and storage but you can also set that up in the tiers of storage to where you have a high performance, high capacity solution. We believe this is the first scale out architecture with sublun tiering in software defined storage. So that's really big. The adaptive optimization. It sets you up for V-Vols, does it? It's very, very complimentary. We've had lots of work where we were heading down that direction for V-Vols and now the performance with this is amazing because you can think of every person has older storage that maybe they have to throw out, maybe they have to retire. What if you could take that competitive array, make that and use it for what it is, just the dumb drives that are underneath it, make that your tier two storage and have a new server with some solid state in it but you virtualize that as one array and now you have a very high speed performance solution utilizing equipment that you already had in there. This is that cost optimization we were talking about. So we got to geek out a little bit here. David, I know you're very excited about V-Vols. You put an exclamation point in your research note the other day. So explain V-Vols to the audience and the connection. Well, the challenge of managing VMware from a storage point of view is that you still have the LUN there. So if you provide any service, you have to provide it to the whole LUN basically. So you want to be able to split that LUN down into much smaller pieces and manage those at a VM level rather than a specific VM level rather than the LUN itself. So that's what V-Vols is about. It's enabling small objects, sub-objects to be attached to the VMware and breaking your LUN down into lots of smaller ones. So more pooling, more granularity. Absolutely, yes. And which is what you were talking about with the left-hand. Sure. Being able to use sub-LUN, sub-LUN tiering underneath that. And having that as the single point of control. Anything you'd add to that, Dale? With the Star Virtual technology, as I said, it's a scale-out technology. So you can scale this up to 16 nodes. And as Craig had said earlier, we've got some new packaging, a real affordable, a package of three with four terabyte solutions each. And that's under $3,000. So all of a sudden you have this insanely price-sensitive solution all the way up to the 50 terabyte solution. So if you can cluster 50 of those together, that's a giant, high-performance solution that was never available before, fully integrated with where VMware and where Microsoft are going with this. So I'm just very excited to be part of this technology. I raised my hand and said I want to be leading this. I want to be driving this because we've got customers that you'll hear from today that are proof points of what HP's been doing in this. And while a lot of companies out there are getting funding to go figure out how to do this, I feel like this has been what we've been doing for 10 years. It's had different names. It'll have a different name probably in a couple years, but we're going to still keep providing that value to customers. All right, yeah, well listen, thanks very much for coming on theCUBE and helping us squint through some of the SDS washing. You guys, like I said, have been there for a while. So congratulations on being in the right place at the right time and really appreciate you coming on. All right, thank you very much. All right, keep it right there, everybody. We're back with our next guest. This is the Software Defined Storage Spotlight. We're here live. This is theCUBE at VMworld 2013. We'll be right back.