 Live from San Francisco, California, it's theCUBE at VMworld 2014. Brought to you by VMware. Cisco, EMC, HP, and Nutanix. Now, here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. Here live in San Francisco, California, this is theCUBE, our flagship program. We go out to the events, we extract the civil noise, I'm John Furrier. Here's my co-host Dave Vellante, our next guest, Michael. Hey, VMware Alliance Manager for HP. Welcome to theCUBE. Thank you. Hot topic is VVOLS. So, obviously storage at the center of all the action in VMware, cloud, mobile, social, sub-defined data center, hybrid cloud, et cetera, et cetera. Moving up the stack. Application-centric theme is certainly here. So, give us the 101 of VVOLS. What is it? Why is it important? What's the key conversation? Sure, and maybe let's start so. So, VMware Virtual Volumes, maybe we kind of start with the motivation behind this new technology. It's been one that's been a buzz. I mean, it's been kind of a technology preview since, I think, 2011 at VMworld was the first time they showed it off. So, it's been this technology that's been a long time in the Cummings. And the basis is what VMware found is that their original architecture for storage, based on these large data storage, these large loans from the storage system, they worked great. It allowed them to consolidate a lot of VMs who have a lot of VM entities on a few storage resources. But the challenge with that is they realized as they kept going and looked at more management and policy-based management, is that there was a disconnect between the language storage spoke, which was at that loan level, and what VMware wanted to speak, which is really at that VM-centric level. So, with VVOLS, what they've done is they've built an architecture from the ground up really to fix that disconnect and provide a one-to-one mapping between the VMware world where that talks VMs and the storage system, which talks loans. So, you'll have, every VM will now have its own volume on the storage system. Okay, so, a lot of people watching will say, oh, I get it. A lot of people go, what the heck is Michael talking about? So, LUN stands for logical unit number. Okay, that doesn't mean anything to have the people out there. So, let's unpack that and really do a one-to-one. So, talk about the storage view of the world versus the VM view of the world and how VVOLS is bringing that together. So, what's a LUN, what's a volume? Let's get down to that level and explain to people. Let's dig down there. So, let's start at the storage, and the way we work at the storage is again, we talk about these volumes or logical entities. So, if you think of your laptop, you have a C drive. You put everything on the C drive, that's your volume of your computer. Some people may have two hard drives, you have a C drive and a D drive. That same concept applies to enterprise storage. We have these entities, these logical entities, where we kind of put all the storage behind it. With large enterprise storage systems, we combine a lot of disks together and create these logical units, and that's where all the data goes. Talking about virtualization with VMware. So, let me just stop you. So, physically, the arrays might be distributed, but you can create a logical unit out of those distributed assets so you don't have to move data. Correct. Right? It's because moving data is a pain in the neck and it's slow and it's expensive. Okay, great. So, that's good. I like C drive, I get it. Okay, good. Why is that a problem? Why do I need something better than that? So, the challenge is again, so we, the storage, specifically at the block level, so we're talking about kind of different protocols of storage, your filers and your block storage. From our standpoint, we know kind of what bits are where, but when you have an operating system, a hypervisor on top of that, that file system knows where everything is. We, the storage layer, all we know is all we see are zeros and ones, essentially. So, it's decrypting that zero and one. Zero and one is what the file system does. And that's where historically, the virtual machine file system from VMware kind of sat on that large entity from the storage system and controlled, kind of understood where everything went. From that level, they put multiple VMs in this one unit and they put them all over the place. We at the storage level don't know what they were putting where. So, if you as a user have, let's say your Microsoft Exchange application and let's say your Oracle application running, then we don't know what bits, kind of what blocks, what pieces of our storage go with which application. Okay, and that's a problem because it creates overhead labor cost, essentially, to try to figure that out and performance tune or move data. And the big problem is, there's not so much data that we don't know that, that's fine, the storage layer provides a lot of advanced data service. So, we do have a lot of capabilities at the storage level. So, things like replication, snapshots, you want to create a copy of your data. Right. So, from our standpoint now, from the VMware standpoint, we want to be able to copy on a per VM basis. But, unfortunately, you're locked into a LUN and so you got to take those services and say, okay, these services apply to this logical use. That's the unit basket of things. But I want that service with that application and I want to drive quality of service. I can't do that without a lot of heavy lifting. Yep, exactly. I mean, the other way we describe it is kind of this big basket of VMs or basket of eggs. And in today's world, if you wanted to paint an egg, you have to paint all the eggs the same color. But let's say you want to paint the eggs different colors. Right, I want one green, one blue, one red for, you know, one white for each of my kids. Exactly, exactly. You know, try to be fair. Okay, great. So, VVOLs essentially will allow you to expose those storage services so that they can be consumed at a virtual machine level. So that then you can align the services with the virtual machine and create quality of service and maybe charge forward of your service provider or an internal service provider. And the result is going to be a lot greater simplicity for my IT operations and better service for my customers. Definitely. That's the bottom line, right? Okay, that's good. Thank you for that description. Why is it taking so long? That's a great question. It has been in the works for a while. I think fundamentally it's completely kind of re-architecting that the storage system just proved to be, I mean, it's a big challenge. Not only that, but VMware does a great job. They have a number of storage partners and everybody approaches storage a little differently. So kind of the original plan of, hey, let's kind of have this one way of doing things. They realize, no, maybe HP3 far does it one way. Storage vendor B does it a little differently. There can't be kind of a one fit, kind of one fit. So there had to be a lot of customizations that I think went into it. And then also, as that VMware stack has gotten more advanced, they realized that vSphere is now kind of creating that fundamental infrastructure of their software-fine data center and user computing cloud. There's a lot of pieces within the VMware stack and now touches and had to make sure it was ready for that. So this is a big win for a block storage provider. So good news for you guys, three par and other parts of your portfolio and others, because generally VMware, as you utilize NFS, as the sort of default, right? So it's, so talk about that a little bit. So does this, I guess my question is, does this level the playing field between block devices and say NAS devices being equal, sort of on an equal playing field? The shorter answer is yes. Historically, the discussion between block and file has been a big discussion with VMware. They created their file system to run on block storage. So that was the first storage they supported. But then as as filers emerge, they realized that they wanted to be able to support that storage architecture. And because Vmware is based on files, that the filer could actually see the individual VMs. They could actually do that decoding we talked about. Like they could pull those eggs out of the basket individually in the past. So when it came to technology like replication, they could do that on a pervy and base it easier than the block storage. Even though fiber channel block storage still the dominant use for vSphere. Now with vVolves, it definitely levels that field, that playing field. VMFS goes out the window, both NFS block will now live on the same storage architecture. So all those advantages kind of go out the window. So we'll all have the same ability to see things on a pervy and basis and at that same level of granularity. Now, you remember, you remember well, I'm sure of VAI and VASA in the last three or four or five years, I guess, let's see. It's been probably four years now. Maybe three or four years with the storage integration wars. Let's call it discussions. It's been a better term. But the issue was that you had to get an SDK from VMware, which they didn't necessarily give to everybody because they had the resources to support that. But anyway, the cartel got it. You know, HP, EMC, NetApp, IBM, the big guys, Dell, a few others. 3PAR actually got it early on. Which was interesting. You guys were a small company, but you got in there. The camel's nose was under that tent somehow. Which I think speaks to the architecture. It's affinity with VMware. But anyway, there was a lot of work being done on integration with VMware APIs. So are we now moving to a sort of a new battleground? I'll call it a new integration battleground where VVol interaction is the new fundamental way to do things? Yeah, that's a great question. VVol's inherently is a framework. I mean, it's this new storage platform that the storage vendors can integrate. It's not a solution. It's not like VA where once you did it, you're done. I mean, once you support the different scuzzy commands, kind of different commands of VA, you're done. Check mark, let's move on. With VVol, this is this fundamental layer where now we can, as a storage vendor, provide all of our unique capabilities through this layer. And those show up in VMware storage policy based management. These policies that they've introduced to control storage. And that will be this ongoing ability to, as storage vendors, kind of innovate, create new features from our side, like our dedupe capabilities. And all the capabilities we've announced recently will now be able to have those shine through this VVol's layer and show up right in vCenter when you're creating VMs and choosing your storage. So there's differentiation is what you're saying. So in other words, the differentiation with VAI was, of course, in the array itself and how things were done. But you're saying it was binary. Once you supported that capability, you had it, checkbox, okay, done. It sounds like this is different. So while it may level the playing field between block and file, are you arguing that it actually creates differentiation between block array companies? I definitely think it has the opportunity to create that differentiation. Well, how so? Let's unpack that. Yeah, again, those, you'll now be able to create these policies which will be based on all the capabilities of storage. And every storage vendor has different capabilities. So from the free part, let's use an example. We have our ASIC that's built in, that's inside the box that provides a lot of our advanced services. D-duplication will be in line in our ASIC. We do things like zero detect to allow you to keep an efficient, thin system. So to kind of pull out zeros and keep your volumes efficient. Well, actually, we're going to be providing that zero detect as one of these capabilities to VMware. And so as when you as a vSphere admin, as a system admin, if you're creating a VM, you can now choose a policy like, hey, I want to make sure this VM has zero detect. Like I want to make sure it remains as efficient as a possible over time. And anytime zeros are written, it gets pulled out. That will be visible by the system admin to vSphere admin. If three-par is the only vendor, for example, that has that capability, then that selection will automatically go to three-par if you have multiple storage in your system. So at this point, both storage admins and now vSphere admin admins, when they'll be able to see kind of all these capabilities and choose what their underlying storage does. So I'm going to build a portfolio of services if I'm a sysadmin or a VM admin based upon the underlying infrastructure and the services that they provide to me. So I'm going to pick and choose. And I might even choose to offer one over the other or maybe I'll have different classes of service depending upon the quality of the implementation. Yeah. And the one thing that we haven't touched on too much is VVOLS also makes all the storage provisioning simpler. That happens automatically behind the scenes. So when you create a VM and choose to apply it to VVOLS, all the, basically the storage work, kind of the setup, basically formatting your C drive, going back to the laptop example, kind of installing this new hard drive. All that provisioning happens automatically. So the storage admin no longer has to provision these big baskets, these big data stores anymore. It just happens seamlessly behind the scenes with the integration we've done. So the vSphere admin will now actually be, when they create a VM, it will also be provisioning storage and doing kind of storage tasks even though they don't know it. So what's, from a VMware perspective, where are we? So it's sort of 2011, I think it was like it showed a little leg. Wow. VVOLS, it's coming. And then yesterday Pat talked about VVOLS is being part of vSAN 2.0 maybe, right? So, but in beta, I mean, it's still a long way to go, it feels like. Let me get the party line. I don't want to speak for VMware. Currently VVOLS is available in the vSphere beta. So if you're a vSphere beta user today, you can run vSphere, it'll have VVOLS in there. If you're underlying a ray that you have and that they announced three storage vendors are on the beta site today, HP3par being one of those, you can use VVOLS today and it's available. You can start playing around with it in your lab. Only three vendors. Three vendors are available, only three vendors today are available. Who are the three vendors? HP3par, and then Dell and NetApp. HP3par, Dell and NetApp. Yes. No EMC. No EMC today. Interesting. Okay. I thought the deck was stacked for EMC. A lot of people think that. Yeah. All right, we'll have to dig into that a little later. All right, well, that's great. I mean, it's interesting. Three par, like I said, is always, even before the HP acquisition, three par was kind of in the club. We did a good job of leveraging our strong, kind of, I mean, service providers, kind of that core cloud customers at the time that VMware was also going after. So we had some very vocal joint customers that helped make sure we were in the door and did a good job. We were the first with VASA, one of the first with VAI. Well, we were here when the first Dell bid came down, John, you remember that? Yeah, that was five years ago. VMware 2010, we were sitting at the desk and boy, things were scrambling. They were all running. Was it 2010 or was it 2011? 2010. Was it 2010? 2010, remember? That was right, 2010. Wow. You've been- Also, a compelling drop right after. Well, that was the prediction, right? When Dell lost out to HP. Bridesmaid. Thanks for sharing the data on- Bridesmaid, come on. That's good, Bridesmaid. You're still good. The compelling was cool. Hey, at least you're at the party. Not looking on the outside. So, VVALS is big. Thanks for sharing the data. Obviously, automation DevOps, this is where the future's going. You know, eliminating all these complexities, manual, provisioning, all that stuff's going to go away. Appreciate it, Michael. This is theCUBE. You're watching live in San Francisco, here at VMworld. Three days of wall-to-wall coverage. We're in the midst of day two, 50 year. We'll be right back after this short break.