 Hi, everybody, we're back. This is Dave Vellante, wikibond.org, with Stu Miniman, my co-host in this segment. Chad Sackett is here, a longtime Cube guest, good friend of the Cube, Chad. Great to see you. Dave, it's my pleasure, as always, man. Stu, it's good to see you, my friend. You know, it's unbelievable, right? John and I have been talking all week. We started theCUBE 2010 at EMC World. We did SAP Sapphire the week right after, and then the big show for us that year was VMworld 2010. It's the best show in town. It really is, it's, you know, we said at that time. Greatest show on Earth, man. We're betting the house on VMware, you know, as a topic, because it is the IT economy. We've obviously spent a lot of time, and as you know, effort, and appreciate, you know, the shout-out that you gave us the other day on the research that we just did. I appreciate the shout-out that your results found. Well, you know, it's all legit, you know. As you know, we do our homework, but Dave and Stu put a lot of time into that, Nick Allen as well, so we're really proud of that work, and at the same time, things are evolving. I guess I want to go back to, it must have been 2009 maybe, we sat in the room, and you chalk-talked the future. Storage, networking, security, obviously compute, management, and the whole deal, and we spent a good four hours in that room. I mean, you know, I was spent after that. But I was just getting started now. You were just getting started. But pretty much everything you laid on us that day is coming true. Yeah. It's really true. I mean, you said storage is going to be invisible. Eventually we're going to get to the network. I mean, the security piece is on and on and on, so. So, you know, I think that that's definitely the story of this year's VMworld, right? The idea of what VMware has done by abstracting the control plane of networking with NSX. You know, prior to that integration to CIRA, having talked with a lot of the CIRA customers, they're super happy with it. But prior to that, it only worked with OpenVSwitch, which meant it was reserved for the customers who were going all in with KVM and with Zen. Now, in vSphere 5.5, it supports the distributed vSwitch and vSphere, which means that idea of network virtualization can be applied to a large swath of customers. And likewise, VMware is doing the same thing for the control plane of storage, which, you know, Stu, that was starting even when you were intimately involved at it, when you were at EMC. Absolutely. Around control abstraction using VASA and the early ideas of VVOLs. And the other thing that's going on is they're disrupting the data services plane by becoming a storage vendor with their own storage stack with vSAN. Yeah, and we're going to talk about that for sure. I want to come back to NSX. Okay. We'll spend some time on that. Go ahead. Go ahead, Stu. So that's just your wheelhouse. Yeah, so on the storage piece, you know, unpack for us a little bit, Chad. You know, we talk about storage becoming invisible, and we'll talk about it in the network space. You know, what is the value of the storage array or the storage stack itself, and how does that play with VMware, especially as we look down to as everybody's showing VVOLs here? So look, the name of the game is hyper-automation in the end. It's not storage. It's not networking. It's not even compute, right? And we talked about in the past that the design and the dream of the Software-Defined Data Center, we used different words for it back, you know, four years ago, but the vision of Joe and all of the parts of the Federation of EMC, VMware, and now the third one, Pivotal, is to try and say, how do we make all the infrastructure in essence invisible? Pivotal even takes it further by just saying, hey, we'll just use Paz and get rid of even all the other structures of service. It used to work for you four years ago. Right. I remember it well. Did you really do that? Yes. So the reason for it is that it sucks when you try to provision something, a workload, whatever the workload may be, and the long tail in the process is touching the physical infrastructure of storage, networking, and compute. Virtualization historically has tackled, and I would call that problem for compute in essence solved. Can there be improvements for compute? Sure, bigger, faster, stronger, right? In storage land, inevitably, we move from the stage of, you talk to the storage person, they provision something to you, to the storage person provisions a pool of something, and then you can automate that and do stuff in vSphere and use it through plugins and automation. Ultimately, though, you would not even want to have that step. You'd want to have the storage advertise its capabilities, and then when the VM gets created, it says, I want out of that catalog of services, this stuff. And that's what VASA and the whole storage policy-based management stuff from vSphere 5051 and now 55 were all about. In networking land, you don't want to have to configure VLANs. You don't want to have to configure firewalls. You want it to be all able to be done programmatically, and the only way to do that is if you can, like we just talked about, with storage and with compute, abstract out the network topology. Yeah, I mean, I really look at it, what we've always said is we need to get rid of that undifferentiated, heavy lifting. So the question I have, there's a lot of startups in this space that have built their products for this new generation, built to say, VM aware if you will, or just simple, simple. And the critique on EMC is that this is legacy equipment, and while it might be integrated and you're updating it, this was still legacy architectures. How does that fit into the new world? So, you know, when you are the leader, everybody will throw stones at you. And occasionally, even as the leader, sometimes we throw stones at others, and I don't like that, right? But I think you might be talking about our friends at, perhaps, Tintree as an example. Well, they are one, they are built for virtual environments, absolutely. And if you take a look at what everyone who is in this space, new players, emerging players, we're trying to today hack at that problem. Tintree too, right? Because there's no constructs at the vSphere layer for VM awareness. What they do in their NAS stack, we do in our NAS stack is to say, aha, a file is an object, a file can be snapped, a file can be replicated. And if we hyper-couple it into vSphere using plugins and extensions, we can then manage and operate on those files, right? Now, again, I'm not saying that our implementation, it's up to the customers to decide about whether EMC's is better or Tintree's is better, and ultimately the customers choose, right? But basically we're all kind of trying to hack at that, because right now VASA, which is the official policy communication vehicle, only operates on data stores, data store unit of granularity, right? VVALS has always been the target of how we would all as an industry do that right. So I would argue that what we showed today about RecoverPoint and the Splitter Driver and being able to do Tivo-like functionality for a VM or replicate for a VM, I would argue we more than hold our own with the competition, but the right answer ultimately is actually to keep going down the path of VVALS and the evolution of VASA so that it can be done correctly and not fake VM awareness, but actually have fundamental VM awareness. So since we started on storage, I got to chime in here. So a couple things. So I asked Pat this morning and his response was essentially, hey, it's all good, these guys are on board, but I'm skeptical. So the- About what? Here's the about what. So as I said, sort of off camera, Microsoft and Oracle have already been grabbing storage function in their narrow little parts of the world, but EMC, NetApp, everybody else, EMC and NetApp are particularly able to find ways to add value and compete very effectively there. VM wears this horizontal player and doing something like V-SAN. Yep. It's software defined, this is the future. I said to Pat, well, don't guys like EMC and NetApp and certainly HP and Atachi and IBM, et cetera, don't they want to do their own software defined? He goes, yes, but they're sort of bought into this. And what do you think about that as EMC? I think, I don't know whether it's right to say it on camera or not. I think that basically as NSX was announced and V-SAN has been announced and everybody in the industry has known that these things are coming, you could hear audibly people's- What the? You're kind of the Cisco of V-SAN, right? So V-SAN's idea of saying, hey, I'm going to glom the storage that's in the server, the DAZ, the flash, the PCIe-based flash and use it as a distributed storage layer is a good idea. It's an idea that is real and innovation is non-containable. As Pat would say, he's a super fan of Andy Grove, right? You know, is his mentor. Andy Grove had a famous quote that basically said, innovation can't be stopped. If the incumbents don't do it, new startups will arrive that will do it, right? That's fair, right? Sam Palmasano as well said, you're going to get commoditized no matter what. The key thing is that it will take some time for V-SAN to mature. The 1.0 target was correctly positioned in the keynotes as use it for non-persistent VDI, use it for test and dev. Customers are slowly starting to grok the idea of, hey, wait a second, this thing by definition has to create multiple copies of the data on multiple servers so its space efficiency is not as good, right? But I think what's going to occur is people are going to start to use it and they're going to dig it. Yeah, they're going to want more. And they're going to want more, which is great, right? Now, from our standpoint, EMC sales reps may not like it, but EMC likes it because you know what? There are portions in the market where we have had great success, taken lots of share, continued to outgrow the competition, but there's other places where we frankly failed to serve properly. And if those customers choose V-SAN, kumbaya. Customer happy, shareholder happy, it's all good, right? V-SAN will expand though, right? And in fact, as a company we embrace the idea of a software-only data service and this is a data plane thing, not a control plane thing. That's why we acquired ScaleIO recently, right? Because we were looking, hey look, V-SAN will be the answer for customers who love VMware and are 100% VMware. And I talked to a big one today who were like, yep, that's us. Likewise, I talked to a huge one that were like, nope, we need an answer that's like V-SAN, but works with KVM, Zen, and Hyper-V, and V-Sphere. Some people like their stacks to have lock-in at one point and they're traded that off and your surveys showed that, right? Others. About half or one to live with that. Right, and they get function and simplicity, right? And V-SAN will be phenomenal at that as people are seeing now, right? I've been using the beta for a long time, so I know it. But the reality of it is that it's going to be a broad kind of ecosystem of traditional storage stacks, embedded into hypervisor storage stacks, ones that are packaged as bring your own server, aka V-SAN and ScaleIO type things, ones that are packaged as we'll give you the server too, Nutanix, SimpliVity. We live in a beautiful, chaotic world. So, Floyer and his piece, the piece that he and Stu did, took a little shot at the cartel and you didn't like that. You shot back and said, no, no. That is absolutely not the case. That's not how we roll. Yeah, it's not how you roll. So how do you roll? The cartel. You roll is what you kill, right? The cartel. So, listen, to be very blunt, I'd be lying if I didn't say that there weren't moments, whereas EMC, we don't get frustrated, that, hey, VMware, you should always work with us, right? Again, it happens more in the field rather than from a headquarters standpoint, right? There's times where VMware gets really grumpy when EMC is supporting Hyper-V or OpenStack in a customer, right? There's times where VMware is really angry that Pivotal runs on AWS. And like the announcement earlier this week was, hey, it works great on vSphere. Like, so think about how weird that is. It's been like running on AWS for a while. Now it runs on vSphere and on VCHS, right? Joe is, I think, Joe Tucci, I think I have an insane amount of respect for that guy. He was wise enough to go, I need to resist the temptation to simplify for own internal purposes and create lock-in from the Paz stack, through the App stack, through the VMware stack, through the EMC stack, and instead say you must all fight for the customer independently. And EMC, you have to pursue it, assuming that VMware isn't a constant. VMware, you must pursue it as if EMC is certainly not a constant. Pivotal, you should pursue it as if neither one of them is a constant. Now, the one thing that I would highlight to everybody who's watching is don't misunderstand what I'm saying. At the same time, whenever one of them is not the best choice for the other, Joe goes, hey, what's up, guys? Who's got this? Yeah, who's got this ball, right? So when VCHS was being stood up and they were looking at alternate storage choices, Pat didn't say you have to use EMC, but he knocked and said, guys, we're looking at different storage choices. You better come in here and if you don't win on your own merits, we'll go with someone else. Now thankfully they did, right? And we made that argument. But you're saying if part of Pat's $50 billion TAM comes out of EMC's hide, well that's EMC's problem, they got to figure out how to shore it up. Yeah, we have to figure out how to compete. Chad, I wonder if you own the global SE force for EMC. In this ever-complicated world, it wasn't easy when you created the V Specialist Force, but it was focused on VMware and they got a lot of weight behind that. There were product managers marking people all with VMware titles inside EMC. In this world of open stack and Hyper-V and KVM, how do you deal with that in the field? So that's a great question, man. So the first observation, just while there is diversity, right, your survey reflects what I tend to find at my customers, right, which is overwhelmingly VMware within the enterprise, use of some KVM's and open stack, you know, where they would have used vSphere or the vCloud suite, a little bit of dabbling in the enterprise, some enterprise customers more than others. Cloud service providers. Cloud service providers far more, right? When we were doing the V Specialist thing, it was an effort to rapidly ramp things up and so we built small focus team, small focus product managers. What's now happened over the last four years is, you know, if you think back, man, like EMC was like a no-show at VMworld's 2006, right? We, our company got the memo focused in, we won the best storage choice for VMware, deepest integration, blah, blah, blah. What's happened makes me very happy now is that's now embedded into the product teams. It no longer requires someone watching it, just happens organically. Good. From a field standpoint, the vSpecialist role, many of those vSpecialists are now leaders of the SE orgs and all sorts of functions. So it's no longer somebody thing, it's now an everybody thing, right? But the vSpecialist mission, which used to be make sure that EMC's the best choice for VMware, has broadened out to really be best choice for the vCloud suite and VMware stack and also OpenStack. To understand and reflect the fact that it's a dynamic open world. So we're already getting the hook and we're going to talk about networking. So we're just going to ignore the hook for now and talk about networking. So NSX. Yes. Awesome. We saw Martins, you know, a little demo up there, but it's not going to be that simple. Why is networking so, so hard and, you know, now remember 2009, you showed us the roadmap. Now we're here, where's it going? So first things first, what he demoed, it is actually that simple if you can constrain a whole bunch of parameters, right? So if you can constrain yourself to every endpoint is a distributed virtual switch or an open V switch. It's like that thermodynamic problem. You guys know what I'm talking about there. So if some assumptions simplify everything. If you can constrain it and say everything is connected to a distributed V switch from VMware or an open V switch from KVM and Zen and you assume that the net physical network layer is a bottomless pit of bandwidth and latency. In other words, you know, that there will never be a contention, you know, at the core networking layer. It actually is really that easy, right? Now that may sound like Chad, those two constraints are stupid. They're not actually that stupid, right? Within the core data center, bandwidth is very easy to apply. It's much easier to say I'll deploy 10 gig E and then go to 40 gig E than it is to hyper design the data center with QOS and manage, you know, customers have demonstrated time and time again that they'll just go from one gig to 10 gig to 40 gig to 100 gig rather than trying to hyper engineer the whole thing inside the data center. In the WAN, different story, right? Also, I mean, it is a true statement to say in a service provider and in most enterprises, 80, 90% of their workloads do finish on a thing that is attached to a distributed virtual switch or a physical switch, right? Now, where it's going to get funky is that obviously I think NSX is perhaps out there in front in terms of SDN land, but they're not alone, you know, and there were lots of partners and we know that Cisco has got some cool stuff that they talked about at Cisco Live and that are coming, right? And you can't, you know, Cisco is an amazing company and they have many beloved customers and CCIEs and CCNAs around the globe that, you know, are going to be very interested to see what Cisco's response is. Yeah, we have Sony and theCube, NCME networks is an interesting play. And you can read all about it online and people's speculation, sure. Yeah, awesome. It's going to be cool, though. I mean, I think one thing that is fun to remember is like innovation is non-stoppable. It's disruptions can't be stopped. They're going to happen no matter where. And in the end, it's fundamentally all good for the customer, whether it's for me or NC, VMware, NSX, whatever. Well, I love, and I said this to Paul Moritz when I first heard his, you know, vision. I said, you guys, VMware is ambitious, you know? If nothing else, it's ambitious, and it's executed on that ambition, so it's awesome to watch. And, you know, stay tuned for next week, September the 4th, speed to lead. There's some exciting stuff coming from EMC. We generally have learned over the years that it's not a good idea to do mega launches and big things during this week, because like you said, this is VMware's show, and it's the greatest show on Earth, right? Yeah, okay, so we'll stay tuned for that. We'll be watching. All right, Chad, thanks very much for coming on theCUBE as always. It's my pleasure, guys. Thank you so much. Keep right there, buddy. We're right back after this quick word.