 brought to you by headline sponsor HP. Here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. We are live in Barcelona for HP Discover 2014. This is theCUBE, our flagship program. When we go out to the events and extract the signal from the noise, I'm John Furrier with Dave Vellante. Next guest is Rob Stretcher, Director of Product Management at HP Storage. Welcome back to theCUBE. Thank you. Great to see you. So storage obviously is always highlighted in their earnings column, always highlighted as an area and storage is not going away. We'd love to talk about storage as that's all we do is talk about storage and how it's changing. So got to get your perspective on where cloud is changing the game. Obviously the show here is cloud, mission critical. Three-par has been a huge success part of HP but now OpenStack is on the radar. I've had people tell me at the last Amazon event, I trust OpenStack over Amazon, mainly because the venture capitalists know what's going on in OpenStack and there's more enterprise action going on OpenStack. Obviously even though Amazon certainly is putting out their own storage for their world. So you got OpenStack, you've got on-premise storage. What's going on? Give us the quick tutorial of what's the current state of storage with cloud data center. Sure. So I think the great thing about OpenStack is that it's got a lot of momentum. Back in Paris, a couple weeks back, there was about 9,000 attendees that are actually contributing into OpenStack, almost as large as a VMworld was here in Barcelona. So if you look at that group of people who are actually building into OpenStack as an open source product and you see how we're able to leverage for that matter from a storage perspective, we actually into our healing OpenStack distributions, we put it into the open source community and that's how it's picked up in healing. So a lot of it is the momentum, openness in giving our customers a lot of choices around how they build their clouds and how they go forward with things like OpenStack. So 9,000 people at the event, that's sizable. It's significant, and a lot of those people are actually writing open source code. So it's not just the people who are traveling there. There's a good number, probably about 80%, where it was their first OpenStack event. And I think that shows a lot of the momentum that's picked up, especially here in Europe. It's always been big in America, again, having choice and looking to build a hybrid cloud and you'll hear a lot about that from HP in general, about how you have on-prem and off-prem and how do you marry those two and a lot of people are looking at OpenStack as being that bridge. So what's the state of OpenStack? Maybe you can help us understand some of the key components and specifically HP's contributions and roles, where are you guys spending it? Sure, so for the Juno release, and it's out there, actually on, they have a webpage that actually tracks who's contributing how much and to what parts of it. HP contributed for the first time was the largest contributor to OpenStack in the Juno release, which is the J release that just went out. And I think that just shows HP's commitment. Meg's talked about it. You'll hear more about that over the course of this entire week, but I think that our big contributions from a storage perspective is really in the Cinder, which is the block storage portion of it. If you think of it Cinder, Cinder Block, kind of the play on words, that's kind of how they try to do it. We're also working for Kilo, which is the K release coming out in April, on things such as Manila. If you think of Manila, what does that sound like? A file folder? It is the file services driver for OpenStack. So you guys did the fiber channel driver for Cinder, right? Correct, we did one that works with our three-par storage and also works iSCSI with our three-par as well as our store virtual products. And then you guys are a main or the main contributor you're saying for Manila? Is that right? We are one of the main contributors for Cinder, along with a consortium of other vendors that are out there. We are working with that same consortium on Manila. Manila is actually pretty much leveraging work that was already done in Cinder and bringing it to the file. So where does Swift fit into this? Does Manila essentially replace that? Does it get subsumed into Swift? It's another access, it's a third access methodology. It's pretty much one of the more popular ones that you hear. If you hear like we were talking Amazon, we kind of contrast Amazon's their API set with Swift. And Swift is kind of that S3 API. Well, Swift is that for a couple of things. And then Cinder would be the elastic block storage analog. Correct, exactly. So how should customers think about them? OpenStack more open, okay? Is it more functional? Is it will it be more functional? What more should we know about them? So I think it's still early on. I think if you look at the competing things that are out there from VMware, from Microsoft, from Citrix, other vendors beyond ourselves who have their own distributions, we have our Healing and OpenStack and our OpenStack Community Edition. There's people like Red Hat, Morantis, and others like Canonical that they're in YouTube that have very full featured, but proprietary or extensions on OpenStack to make it easier. I think OpenStack itself alone is still working through the kinks of getting to be enterprise ready. I think if you're going to go down the OpenStack route, a lot of our customers have actually hired people who are contributing back into the community because you need that developer mentality to really get OpenStack going full throttle. A lot of people are confused about the business model with OpenStack. They hear open and they think, oh, it's like a duke, nobody's going to make any money, it's all going to be free. But can you talk about the business model? How you guys, because you really don't care if there are different distributions. Obviously you want Helion to do well. Right, exactly. There's a different from a storage standpoint. You want it to be, you want your products, your software licenses, your hardware products to be compatible with anything that's out there. So can you talk more about the business model? Sure. For us, we see a lot of our customers going towards OpenStack and they're looking at how do they have choice at a hypervisor and at kind of that management control plane? If you think of OpenStack, they have the word software defined storage in there and they have the word of software defined control plane which is the manageability features that is done through other things like Nova which is compute and that's the driver that does being able to look at the processor and understand how the processor is being utilized and do multi-threading and things of that nature and high availability in concert with how it works with our Cinder block drivers. So I think when we look at the business model, we want to be there for our customers if they're going that way. And we see that in a lot of our service provider customers, especially here in Amia that in Europe, they're really looking at if not implementing clouds based on this. At OpenStack in Paris a couple of weeks ago, I was with one of our Japanese service provider customers and they were really talking about how they were moving. They were back on a earlier release of OpenStack on the F release was called Folsom but they were looking at how do they go and I think up until recently upgrading was a problem. So from a business model for us, what we see is we're interacting with those early customers, early adopters to build in the different features that frankly VMware has there today but VMware costs you money and I think it's a balance between open, truly open and going and just downloading the distribution off of the OpenStack.org website or and building your own or going to somebody like a Helian OpenStack and getting a fully supported version or Red Hat or Morantis or one of those or are you going to go and stick with VMware and I think it really has to do with the business model of our customers and their operational stance and what their cost model looks like. And where the leverage is, right? When you talk to VMware, you're at VMware and they say, yeah we love OpenStack, we're all in, blah blah blah but customers like you said want choice, you actually use the term hypervisor choice. Choice to me means, you know, pricing leverage, you know, pricing power options. So, I mean, essentially OpenStack, VMware, Amazon, these are competing entities largely. I mean, yes, there's complementary aspects and certainly people are going to use all three but OpenStack is largely an alternative to those, is it not? Yeah, and I think, you know, and VMware was at OpenStack in Paris along with several others who would be considered competing but I think you look at Amazon as kind of the outlier as closed and proprietary on their gear, for that matter. They're doing some things to open it up a little bit but I kind of look at is VMware has a lot of the full feature, the richness out there today where OpenStack is going, they want to get there based on the KVM, the kernel virtualization hypervisor that's in Linux and if you look at it longer term, you know, VMware is talking about how you'll use vCenter to control OpenStack and be that management control plane and I think that's where interesting things will happen is that management layer and it would be interesting to see how it all shifts from that. Well, what can we learn from other open source? I mean, you remember the Linux Unix discussions, oh, Linux is so much more functional, you know, those days are gone so project 10 years out, why wouldn't OpenStack have the same rich functionality as Amazon or VMware? I think it will, I think, you know, honestly, it's good for VMware customers as well as people looking at OpenStack because they think it's pushing innovation. If I look at where we want to take it from a helium perspective and what we're doing on our storage side, it's about driving the enabling technology but that just means that VMware has to go that much faster to stay ahead with the richness and I think it's, you know, what we're seeing, you know, also that we've learned in the hyper-converged space is it comes back to simplicity and I think right now OpenStack is not simple and it comes down to a customer that can't hire OpenStack-based engineers is not going to be able to do OpenStack very well. They're not going to be able to keep up with the distributions, the packages, the updates, they're not going to be in the stream, it's going to be tough for them. They're going to look to a vendor like ourselves with our helium OpenStack or to somebody like a Red Hat or others to really keep them up to date and to bring that simplicity. But still, the feature richness is going to come from people innovating around the ecosystem and I think there's a play there for VMware too, longer term and I think it all depends on where they go with ESX-I and vSphere and how do they really embrace it or do they not and I think they do have a play there though. The question that I want to ask you is on OpenStack Europe versus the US, obviously the Paris show, not well attended by standards in the US, obviously the Silicon Valley event was learned just through with the industry was packed house, Cube was there, we had a great time. But it's a travel, it's a pretty big travel but it's a different focus, I've heard people say a lot of telco conversations. What's different, what happened in Paris and what's happening here that's different than North America? Because all I was just hearing about the same, enterprise here in Europe is a little bit different. What are you seeing as different conversations? Yeah, the conversations do, especially in the cloud space tend to be different in Europe than they are in the States, a lot of telco conversations and you would kind of look at it as smaller telco or clouds and I think a lot more of them. It's more fragmented market over here because you have so many countries with so many different laws even though the EU permeates around but they still have different pockets of clouds that have allowed people to uptake but I think what they have learned is that adopting things like OpenStack and moving rapidly towards an open architecture can help them on their cost curve and I think that's where OpenStack in Paris was a lot more geared towards getting work done on the standard itself versus being the trade show aspect of it. Although, and a lot of nuts and bolts being talked about, a lot of nuts and bolts being turned and talked about and I think that a lot of people are figuring out, okay, this is the best way for me to proceed forward with that as well. So a lot of people have questioned the progress of OpenStack, obviously there's pros and cons, consolidation, you're seeing cloud scaling was acquired by EMC, Randy's company, some say in a yard sale. I've known, we've known Randy, I guess he's been accused many times, always dynamic that I talk to. Obviously other acquisitions you saw, Josh McEntee go to Piston, he's the founder, so there's a lot of stuff going on behind the scenes. HP, Eucalyptus, Eucalyptus, Barton, so consolidation is always good, you're just going to see some, I think some good progress here, so I want to get your take, not so much what's going on in the Scuttlebutt in the industry, we have a good handle on that, happy to report that another time, but inside HP, I'm almost imagining as if I'm inside HP that similar conversations are happening inside with senior management and down in the trenches. Obviously cloud group is well funded, we've seen them at all the events, Montee's awesome, he's on the board, we have two HP board members up there, awesome people interviewed on theCUBE as well, so big bet on OpenStand, but you're in the storage group. So have the Fyftons talking to each other, what is the conversation inside HP, and tell us the truth. Oh, that's great, and I think it's very accurate. A lot of my team's work from a software-defined storage perspective is working across those groups, be it with our cloud team, or be it with our converge systems team, or be it with the server team straight out, and I think what's nice about that is really the Fyftons have been breaking down, and I think what we're trying to do is really coordinate across the different business units, so that we can make sure that when we're building something for OpenStack, there's a business reason and justification that the cloud group sees out of it as well as ourselves, and that we're doing the right things at the right time. So we're helping them on the standards, they're helping us with the standards, because they see it at a much broader level of where that entire OpenStack business is moving. We're there to kind of help support and make sure that as one HP we can get there, and I think that's the important part. So given that, this code leaderboard issue always comes up, or number one, Red Hat might debate that, and say no, they're number one, they only work on patches, what defines number one? I mean, give us a taste, I mean, are you including every single code contribution, including QA, or is it core code? I mean, what defines that? So I was including everything over the top, and I think that to your point, when you look at the top four, we're all in about the same space from that perspective, between us, Red Hat, IBM, and some of the others are out there really contributing, and I think it just speaks to it being a great way for us to move that standard forward, and I don't get as hung up on the leaderboard as the progress we're making with our customers, and seeing them implement it, and actually leverage the technology. It's really more of pimping up the fact that you guys are big contributors. Not an insignificant player at all. That's me, the main purpose. That is nail on the head right there. I wonder if we can talk about each software defined strategy. It's everybody's software defined, everything these days, and you're an expert, you live the software defined life, according to your LinkedIn profile, I love it. But so you have a lot of different strategies out there, you're seeing EFCs there, okay, we're going to separate the control plane from the data plane, it's actually putting in a layer so they can bridge the old with the new. Okay, cool, we get it. You got NetApp saying, hey, we're going the whole thing. Here it is, we're nuts. Big risk, Hail Mary, big play, okay, got it. What's HP's strategy around software defined? I wonder if you could sort of crisply articulate that for our audience. Absolutely, and I think it comes back to, we do believe in a control plane or a management plane and data services. And as you would hear David Scott talk about it, we really look at it as being unified data services across our platforms, not just software defined, but across the service refined or SLA oriented ones. And our cost optimized is really where we see software defined fitting for people who want to go wide. If you want to go deep, we have our three par, best price per square meter of data center space that you can get, steep and deep, when you want to scale out and scale up, that's really going to be cost optimized and software defined. And we look at that from the data services layers really being our store virtual and our store once VSA product set, where we look at it from a controller management plane is we're going to play with Microsoft's, we're going to play with VMware's control planes. They have their vSphere control plane, Microsoft has their SCVMM control plane or what they would claim is their control plane and we're going to play with OpenStack. We kind of look at the interface for that is really going to be one view in the long term and how we go to market as one HP is leveraging the power of one view to interact in all of those different control plane areas. So, okay, so then my follow up on that is related to the three par piece. So three par, scale up, scale out. I mean essentially is really what you're talking about here. Do you see over time the sort of three par functionality? I mean, doing stuff in Cinder, that's software defined. Do you see that functionality trickling into the software defined umbrella? Absolutely, we look at bringing those services, for instance, our adaptive optimization which was doing the automated tiering at a trunclet level underneath the hood inside three par. We brought that over about a year and a half ago to our store virtual platform and now we're doing even more to bridge that that divide, I guess you could say, between that SLA optimized and the cost optimized, bringing three par and store virtual close together. We'll have some announcements here in Barcelona later today around some of our functionality bridging store wants and three par as well. And that's going to be really exciting and we kind of look at that as a way to get out there and bridge the gap between what had been quote unquote physically constrained and how do you bring that broader for people who may not need a three par but they want that full rich set of features. Absolutely. And I think another area of confusion and software defined is people say, well, it's all software so it must be software defined but I think of it as more than that. I mean, and I wonder if you could weigh in but the ability to through an API programmatically provision infrastructure and deliver an SLA and change that SLA through code and automate that. I mean, that to me is software defined. Is that a fair sort of way to think about it and wonder if you could add some color to that? I think that's a great way of breaking it down for the customers. And a lot of what we talk about is how do you have REST based APIs that are published open APIs so that if you're using OpenStack or using VMware you know that there's going to be a way to do integrations or if you are out there and want to write directly against those APIs because you have a particular application or business need that needs to say take a snapshot at a certain time during the day at a certain interval we provide that out to our customers in that REST based API which is using web based technology. I think that what you'll find is that we look at the control plane as being a place that we need to be able to play in a lot of different environments and I think that is really why we've taken that more open approach to the control plane versus saying hey you have to buy our control plane and our control plane only. So software companies, Rob, they always want to commoditize hardware, right? That's what good software companies do. It's just hardware. Put the value into our software stack. So you're sort of used to that with Microsoft, VMware now with vSAN is taking a similar path but storage has always been really hard and that's been sort of a heat shield against the commoditization trend. With software defined does that change? How do storage companies preserve margins in this sort of software defined world? Is it just adding more function? Is it making it up in volume? I wonder if you could talk about that. Well, I think it's a good, I'll give you kind of an example. If you look at our three-part development team the vast majority of more software engineers. Three-part is software defined inside the box. It's just hardware enabled and I think it's the difference that there are going to be certain things that you need hardware for to actually increase the performance of that or do special features and things of that nature that just can't be done based on the process. We have a long relationship with Intel and we do some co-development with them and we look at how do we leverage those but at a certain point, the ASIC that's in the three-part does some special stuff that you really just can't get at scale to maximize that data center space. Like you can with a three-part system. Software will have the features but you know, it may not scale to the same way that a three-part system can from a density perspective especially. I got to ask the R&D question. Where is it developed? Inside the storage group, cloud group and can you comment, we're seeing some comments on the crowd chat in the crowd about HP Labs not having anything under the table with OpenStack. No mention of OpenStack and HP Labs. So is HP Labs doing any R&D? Are you guys doing any? So you can kind of look at we're all being done in the open. It's all being done in the open from a perspective that there is innovation going on at HP Labs but that's usually a further out type of approach where they're doing things now that will show up five years from now in productization. Where? Specifically at OpenStack. Yeah, stuff that can be leveraged into OpenStack. So you know of stuff going on at HP Labs that's OpenStack related. There is stuff in HP Labs that's OpenStack related. Okay, that's good. We're going to get that out there. So folks out there, HP Labs is working on stuff. They work on stuff without being specific about it. But usually HP Labs has got the propeller heads that do what's a really cutting edge, really far out there. Probably imagine around power and cooling, software weirdness, memsters, all that stuff. Yeah, and I think you look at it and how do you deal with, because at the end of the day, software still sits on hardware. I don't care if it's a server or a purpose-built three-part array, you still have software on hardware. And we happen to be a fully, full-stack integrated company where we know what's going on at the hardware level. So what about OpenCompute? Obviously we cover that show as well. We love OpenCompute. It's got the same ethos as OpenSource. Some say there's really not a lot of meat in the bone there, only for high-end dudes doing their stuff at large scale, whether it's on build your own telcos. What is the status there? Because obviously it seems like a perfect fit with OpenStack. Yeah, I look at OpenCompute as being, it has to make sense. And I think it goes back to, are you going to build your own? Or do you have a core business that if you go and build your own servers, that that's going to lower your operational costs or give you an operational advantage over your competitor. Those are people who go to OpenCompute. I think for the vast majority of people buying a server at cost and seeing the cost continually coming down from that perspective as the processing power goes up will be great. And I think OpenCompute kind of will play a role but not as a significant role as OpenStack. Rob, really appreciate you coming on. Quickly get you the final word in. What's the big announcements on your Converge and your software defined storage? Give us a taste of what you guys are talking about. Sure, so we were talking about hyperconverge. We launched our CS200 families where we announced our store virtual and our Evo line and we'll be talking about announcing and shipping our store virtual CS200 later today. Okay, we are here live on theCUBE here in Barcelona in Spain, covering HP's European customer event. Annual show here in Europe. Obviously great attendance, record attendance again. And we are here inside theCUBE. Getting all the data, sharing that with you. Join the conversation on CrowdChat, crowdchat.net slash HP Discover. Log in, be part of the on the record conversation. We'll field some questions from there and be part of the crowd and join theCUBE, join HP Discover. We'll be right back after this short break.