 Live from San Francisco, extracting the signal from the noise, it's theCUBE, covering VMworld 2015. Brought to you by VMware and its ecosystem sponsors. Now your host, Stu Miniman. Welcome back to VMworld 2015. This is SiliconANGLE TV's live wall to wall coverage of the event here from Moscone North. I'm Stu Miniman with wiggybond.com. We've been doing a bunch of segments digging into this hyper-converged infrastructure on the wiggybond side. We also call the general trend server SAN. Really excited to have a panel to help dig into the reality of what's happening in this space. I'll start on the left of the panel here. We've got Chris Miller, who's the principal architect from Rolta Advisex, a partner of VMware's and two of the team working on the VSAN and solution set, Gaetan Castelline, senior director of product management with VMware. And we've also got Raj Yavatkar, fellow and chief architect of VMware. I hear kind of the CTO for some of the solutions putting in gentlemen. Thank you all for the first time for all of you on theCUBE. Thank you for joining us. Thank you. All right, thanks Stu. All right, so what we really try to do is there's so much going. I mean, so many sessions people are learning about what's happening, new announcements at the show. So we need to give a peek inside this. But before we dig in the products, get a little bit about your roles. Chris, talk to us a little bit about Rolta Advisex and your role there. Sure, so my name is Chris Miller with Rolta Advisex. I've been with the company now for about four years. My primary responsibilities are supporting our enterprise customers with solving problems with technology. And as a principal, I might have a stake in actually enabling our team and making sure that we understand the hyper-converge infrastructure platform from VMware. All right, Gaetan? All right, so I'm Gaetan Castelline. I run a product marketing team on software-defined storage at VMware. So we take care of virtual sand, vivans, and we need to hold software-defined storage stacks from a VMware standpoint. All right, and Raj? And I lead the so-called Evo A-Study C program, responsible for strategy, architecture, product management. And that's our attempt to show what's the easiest way to do private clouds using hyper-converge infrastructure. Yeah, a quick question, Raj. My understanding is that the team that puts together kind of the Evo packaging is a separate organization, even though they work closely, of course, with V-Sand. Is that true? Can you share how big that group is? A number of people is about 50-plus. Okay, yeah. Okay, great. So, Chris, I'm going to start with you. You know, what does this whole discussion of hyper-converged mean to you, mean to your clients? You know, what kind of brought you into that whole space? Yeah, so it's been a disruptive market lately. We've seen over the past five years new storage technologies emerging, and we've heard our customers all saying the same thing, which is that we have to reduce our costs, we have to do more with less, we have to be more efficient, more scalable. And so, now that we've got this hyper-converged technology that's sort of emerging in the marketplace, we've seen more interest in this area of technology than we've seen in most of the past five years. So for me personally, it's a way for me to kind of get on the cutting edge of technology, understand how we can help solve problems. For customers, it's solving multiple problems at once. All right, and just curious, Rolta Advise X, was this just an extension of what you were already doing with VMware? Did you have an existing storage practice? How did that get into it? So we actually, we grew up in storage consulting for a number of years. The company's been around since 1975, and we've been evolving with technology as well. So for us, this was just a natural extension of our data center practice. All right, and Gaitan, let's talk about, I mean, the storage is obviously a critical component of everything we've been doing at VMware. You know, it's evident at this show how many of the partners and how many of the sessions can consider the storage component of it. So help bring us into, you know, how hyper-converged fits into the discussion. Yeah, so you know, for us, hyper-converged is a major trend in data center infrastructure. It's really a fundamentally better building block on which to build as DDC. If you look at the way people have been doing infrastructure in the past, it's been about building physical silos. I have my compute silo, my network silo, my storage silo, and the SDDC layered on top of that makes the physical infrastructure better, easier to consume, more dynamic. But you're still dealing with physical clutter underneath the SDDC. What HCI is about is about converging all of those functions and really running that as software on X86. And by doing so, you're eliminating the physical silos. You're making the infrastructure a lot more simple, scalable, high performance and cost effective, and we think it becomes a much better, a fundamentally better building block on which to deploy the SDDC. All right, so Raj, you know, maybe you could help the Evo SDDC, one of the new announcements here. There's also the Evo SDDC manager that goes into it. You know, the thing I've been looking at for many years, even before I was an analyst, is, you know, we want to build things at rack scale. We want to have, you know, pools of resources. We want to get out of the traditional silos that we're doing. So maybe can you explain a little bit that Evo SDDC and how that, you know, how does that fit into hyperconverging? What does it deliver? So Evo SDDC is really a story about integration. Like you said, there are silos across different organizations. So what we're trying to do is that, provide a way by which people can bring together a private cloud at the agility and cost and availability of the public cloud economics we can bring to the enterprises. So SDDC manager is an integration point. It integrates all of the VMware stack, hardware stack, including firmware BIOS with a very prescribed configuration so we can deploy private cloud in less than two hours. Yeah, so Raj, okay, let's poke on that a little bit. So if I look at cloud, I want to spin it up fast. It wants to be easy. Operationally, it's totally different from the way I manage my infrastructure. And by the way, if I want to use less of it, it's easy being to turn it off and stop paying for. So I've definitely seen the whole hyperconverged discussion. I've talked to customers that say, you know, hey, is this, are you like a cloud now? So, you know, how does this kind of compare contrast versus what I think of the public cloud? So I think, excellent question because that's the first question people ask when we are trying it out. One of the things we found out is that when you have one or more racks and you scale to capacity which is multiple racks. If you could present that as a single pool of capacity across computer storage networks, you can start carving it out into domains, workload domains and allocate capacity based on availability, performance, security. And that's exactly what we do when we go to the public cloud. So idea is to mimic that to same extent but also keep it under the control of the IT from compliance and security perspective. And Stu, if I can add something, you know, yes, we're seeing a lot of adoption with the public cloud. I think the reason for that is the simplicity, like the speed of deployment and how easy it is to consume the public cloud. The disadvantage of the public cloud is economics. Like it ends up being expensive, especially as you add capacity into that public cloud. I think the objective of Vivo ZDC is to enable customers to deploy a private cloud almost as easily as you will consume a public cloud but with much better economics because it's your own private infrastructure. Sure, I mean, there's some places where, you know, my on-prem stuff isn't going to compete. If I just have a resource, I want to spin it up for a couple of hours, I'm probably not going to spend the time to do it. So, you know, we always like to say it is forces for courses but the hybrid cloud discussion has definitely been involving. Chris, what do you want to pull in? You know, what's this conversation you're having? Does cloud mean things to your clients today? You know, everybody, I mean, you know, I go to a room, there's two types of people. People that are using AWS and the people that don't realize that they're using AWS and we're all using Microsoft and you know, I mean, most customers are using VMware. I mean, you're dominant in the data center today. So, where are your customers with this discussion? Yeah, no, it's interesting. So, you can walk into a room full of 10 people and ask them all to define cloud and you'll get 10 different answers. And so, in a lot of cases, what we see from an executive level is focus on software as a service, the ultimate agility and where that makes sense. Infrastructure as a service is definitely a goal and a desire but to echo some of the sentiments from earlier, you know, it's really good for transient workloads, things that I'm going to pull back because that's where I can get the economy. Internally, there's still that big challenge which is how do I provide this self-service IT that's agile, looks the way the public cloud does. And so, this discussion is coming up more frequently. It's consumed all of my time for the past 12 months with my customers and that's why I'm really excited about the hyper-converging infrastructure. It's finally the solution. And I'm curious, you know, is this the storage people looking for something different? Are there projects that are bringing in, you know, operationally and use case-wise? You know, what are you seeing? Why are people looking at VSAN? And I'm curious what you can share about your customer's adoption of VSAN. Yeah, absolutely, so a couple of things. We're seeing a lot of this being driven at the director level and above. And most of it's because if you break down the typical organization's IT budget, the biggest cost is labor. The next biggest cost is storage. And if you look at the labor component, a lot of that is storage. So when you look at what VSAN can bring from a scalability perspective in terms of how you actually acquire new assets, there's just the overall cost savings, it's being driven at that level. In some cases, we're having storage teams ask about it, the folks who are more progressive. In a lot of cases, it's the VM where it's being bringing that knowledge to the table. But I would say most of it is coming at the executive and director level. I would add one more thing to that, is that people don't realize it. People who have traditional storage arrays, their workloads have already started migrating to storage virtualization. So it's sort of reactive to them to find out the workloads are no longer owned by them as storage policy owners. All right, so Gaitan, maybe we talk about, you know, your interactions with clients. When the 1.0 product came out, we thought it was really interesting, but there were a lot of checkboxes that we were hoping that we would check. 6.0, you know, check some of the major gaps, you know, no offense, but we said to get it from kind of a test environment to a more production-ready environment. 6.1 has a lot more features to walk us through, you know, you had a lot of customers that were running even the 1.0 product, so what's that discussion been with customers and where are they in regards to the roadmaps? Absolutely, it's a great question. So 1.0, which was actually called 5.5 for alignment with vSphere, was released about a year and a half ago, and we've always had a lot of confidence in the platform, right, excellent levels of performance IOPS will bring in the data close to the compute, we're leveraging SSDs, flash, and that allows us to really deliver tremendous amounts of performance and scalability. What we were missing initially were potentially some of the data services, right, high performance snapshots, synchronous application, things like that, and then we were also, I think, a little bit cautious from a VMX standpoint. It was a 1.0 product, we didn't want to lose any data, and so we guided customers to test and dev use cases, and actually we'll only have successful, because some customers, despite that guidance, actually started putting Oracle databases, Exchange, and so on on the platform, but even before the second release, even before 6.0 came out. We released 6.0 about six months ago, and this week we're announcing 6.1. 6.1 is another big step in terms of having enterprise-class availability, stretch clusters, synchronous replication, 5-minute RPO for racing replication, better management with virilized operations integration, and VME support, so a lot of good new things in the 6.1 product. And then we're also pre-announcing a beta. We're announcing a beta with some very high-demand features, DDoop, Erasure Coding, and software checksums, right, so that's not available yet, but we are introducing, we're doing a public announcement for the beta, so customers will be able to test those things out really soon. Great, Raj, do you want to help fill out the rest of the announcement suite that was made this week? Yeah, so the EOS2DC is a solution available soon from our partners, and that allows you to order a complete software-defined data center as a pre-integrated, pre-image, pre-installed. You bring it in, cable, power, connect, and in less than two hours, you have the first VMs running. Then we also automate the lifecycle management. So you don't have to now invest in different tools for physical infrastructure management, logical infrastructure management. All of it integrated into a single manager called EOS2DC manager that also automates lifecycle management. All right, so Chris, I want to come back to you. You know, the technology itself, when it comes to convergence, you know, is the first piece. The go-to-market, you know, how do I incent the partners? How do you guys make sure you add value into it if something's, you know, completely done? Wait, am I just, you know, selling a license? And the packages, I mean, it usually takes time. I was involved in the early converged infrastructure, you know, and it usually took 12 to 18 months to kind of sort these out. And I'm curious from your standpoint, there's VSAN, there's the VSAN Ready Nodes, there's the EVO family, what's your experience been, and you know, how do you write VMware? And they're not paying attention, tell us the truth. Yeah, so, you know, from a partner perspective, you know, especially for us where we've had established relationships with storage vendors and the products that are out there, there's a comfort zone there. And I think that folks are nervous about, you know, taking the next leap. But you know, what I always tell our sales folks and our pre-sales folks is look, if you're not having the conversation, somebody else is. And you're going to lose account control if you can't bring that new technology to bear. And so we've adopted that, well, we've been pretty good about, you know, shifting when the market is moving in a certain direction. And I think that, you know, when you have an opportunity to save your customer money and to move them into a new paradigm, you should take advantage of that. So we've not seen too much pushback from just a pure VMware perspective. You know, it's the simplest product out there. I check a box, I have a storage pool. And so we've worked with numerous other platforms in the past and they all have their strengths and things, but you know, from just a pure simplicity perspective, the fact that I can walk into a customer, drop a cluster on the floor, click a button, and I have storage that performs with very little maintenance. That's been a very big success for us. We've been very happy with what VMware is doing. So which deployment methodology are you doing today? I talked to a couple of customers earlier on camera and they were, you know, I got VSAN and I follow the HCL and, you know, it worked, but there were a couple of tweaks and the OEM flashcard versus the other one and firmware versions. I mean, we know the horrible compatibility list sometimes is what we have to do, not just for the industry in general, not trying to pick on VMware, but getting that stack right. I mean, I spent many years in an interoperability lab and getting that stack right and making sure it all works together. I mean, that's some of the value that the partner community and VMware are putting. So what is, are you guys doing that for them? Are you doing an EVO package of VSAN ready or just spending your own VSAN? So frankly, the burden's on VMware, right? So for us, we consume the HCL and it's constantly being updated. We have that information handy. So the way we approach ready nodes is we use those as a reference architecture starting point. In a lot of cases, you know, a ready node might fit a certain profile for VDI, for example, but part of what we do as a partner and our value add for our customers is actually measuring workloads and determining what the very specific requirements are to build that design. So we'll start with a ready node. We may tweak it to get better capacity, but what we're really looking for is that sweet spot to get you what your business requirements dictate at the right price. And Stu, if I can jump in here, right? So we're providing customers choice, right? They can do the VSAN ready node, which can be customized, gives them a lot of flexibility. They can do EVO RAID, which is a turnkey appliance, and now they can do EVO SDDC to build out a full SDDC. So they have those options. The benefit of the ready node is that it can be customized. It can really be configured to meet those specific needs. But you're right, the day zero experience is not as good as with the turnkey appliance, but we're very focused on making that experience better. We're coming with software to really automate the configuration, and that's one of the key priorities for us. Absolutely, there's always that balance between flexibility and making sure it's rock solid. I mean, it's great if I can have any car that I want as long as it's black, right? That works, it comes off the production line, it's easy to do. Chris, the other thing I wanted to ask, and we'll let the other guys jump in, when customers are looking at this, is it, they're looking at a lot of hyper-converged solutions, or is it vSAN compared to kind of traditional three-tiered architecture? Where is kind of the competitive environment for this today? Sure, I mean, I think it's a little bit of both, right? In some cases, that traditional architecture has grown too unwieldy and too expensive, so they're looking to hyper-converge to solve the problem. And in other cases, they've been exploring the hyper-converged market, and now that VMware has made its entry there, they're saying, hey, if I'm going to do my due diligence, I need to investigate it. But I want to come back to, just for a brief moment, around the HCL and the workloads. So the ready nodes, the way that vSAN is leveraging Flash and the performance that we're seeing, I'll give you an example, we had a customer that was doing a POC with vSAN, and during the POC where they had some of their lighter workloads running, they had a problem with their SAP database, and so they needed to do a restore, and the target that was there was the POC node for vSAN. They restored their SAP database and brought it up and running, and to their surprise, it ran much faster, lower latency, better performance than they saw in their production environment. And that, right away, finished the POC. So even though we have these nodes and we have these workload requirements, there's a safety buffer there, and I think that having been a storage admin myself, we like to just worry about capacity. We don't like to have to worry about the IO. And the fact that we have so much excess IO, so much performance capability in the product, it lets us sort of manage storage the way we want to. All right. If I may jump in the competitive environment you mentioned, I'm a user of vSAN, and I use a subset of vSAN that denotes to deliver hyperconvigent infrastructure solution. The biggest advantage of vSAN approach is it's integrated into the hypervisor. So you get performance, scalability, but more importantly, you get automation of vSAN to management. Our VMware customers love vSAN, they're used to it. This is just integrated. You don't have to go to yet another tool and try to configure storage. It just comes out, you get the pull of storage capacity. I find it very convenient. Yeah, so just to add to what Raj was just saying, the whole point of hyperconvigence is to make storage go away. You don't want to have to manage it. vSAN is completely integrated with the assets, with vSAN, you're not managing a separate product. Compared to other solutions, we have about 2X the performance at half the cost. So you have full integration, 2X performance at half the cost. That's a pretty good equation in terms of competing with those other solutions. We know there's so much cost. Chris, you brought it up early. The operational aspects of dealing with storage, we want to make it invisible, we want to make it simple. I mean, simple is the thing, and I look at all this space, it's simplicity, simplicity, simplicity. Chris, please. Yeah, to add to that, I think what's generating the buzz specifically was vSAN's half the story for hyperconverge. So one major difference between what VMware is doing and some of the competitors in the hyperconverge segment is VMware can bring not only the storage, but the networking to the picture. And when you look at ERO-SEDC, I like to think of it as push button efficiency. I can drop blocks on a floor, press a button, and the entire stack is stood up. And now I can not only hyperconverge my storage and get the operational benefits, but I can do the same with networking. That is very strong and really unique in the marketplace. Okay, great. So we're running low on time when it could be each an opportunity, just, Chris, maybe start with you. Experiences so far at VMworld, maybe your key takeaways. Yeah, it's been great. So VMware has had some initiatives to focus on enterprise readiness, making sure that products are scalable, that things are all sort of unified in terms of management. And everything I've seen here has shown me that VMware is meeting that goal. The products are ensuring, the feature set's great, the cloud roadmap's great, I'm thrilled to be here. I'm curious, anything specifically that's been on either your wish list or your customer wish list that you'd want to call out from the recent announcements or any white space that you're hoping to see going forward? Yeah, so I think that the public cloud services, there have been a number of additional services that our customers, and we've been asking for. And when you look at the roadmap, it's spot on, so I think we're good there. All right, Gaitan, I mean, you walked through a number of the announcements. Key thing you want people taking away, looking at the vSAN solution set this week. Right, so if I look at vSAN, I think it's really about adoption, right? We're seeing customers asking about vSAN, we're seeing a lot of momentum. It is happening, and we're very excited about that momentum that we're seeing, the level of excitement that customers have. As Chris was mentioning, the roadmap is resonating very well. So we'll feel pretty good about the future for vSAN. Yeah, I think people underestimate how important this is to kind of the direction of what VMware, not just in the storage team, but it's a critical component of what you guys are driving. And if people thought we were going to have a million customers overnight, they weren't listening to you, we're looking at your forecast. I hear, according to your forecast, you're doing above what expectation was. We are, and we actually announced publicly that we have more than 2,000 customers, and that makes us, according to our estimates, the number one HCI vendor out there, right? So we think we take a leadership position in terms of customer adoption in nature. All right, so Raj, you get the last word. Evo SDDC, expecting it early 2016, right? Yeah. Some of the partners showing off some of the stuff already. What should we be looking for? I think you have seen convergence first structure, you have seen hyperconvergence first structure. I almost went to a start a few years back, but what I took away from this VMware world is VMware has introduced now enterprise class hyperconvergence first structure, which also lets you integrate networking, as Chris said. And once you do that, I think everything is XRD6 hypervisor, and that's what the take away is. All right, well gentlemen, thank you so much for helping us unpack some of what's going on in the hyperconverged infrastructure space. Thank you for watching. This is theCUBE. We'll be right back with lots more coverage here. SiliconANGLE TV, VMworld 2015. Thanks for watching.