[music] 00:03 Steve Mcdowell: So welcome to the DataCentric Podcast hosted by Moor Insights & Strategies Senior Technology Analyst Steve McDowell, I'm Steve and Matt Kimball. Say Hi Matt. 00:12 Matt Kimball: Hi Steve, how are you? 00:14 SM: I'm doing well. Surviving the forced isolation? 00:19 MK: Yes. I'm waiting for the zombies to approach my back door, but surviving otherwise. 00:24 SM: Zombies. Hey, so Matt I'm very excited. Today we're gonna talk about HCI and specifically what VMware is doing in that space, and when we look at HCI, VMware's vSAN is the dominant product in the market, it comprises somewhere north of 40% of all HCI installations I think as of the fourth quarter numbers. And continues to grow at about 25-30% clip year-over-year again at least as the latest numbers came out of IDC. There are more than 20,000 vSAN customers and I don't know how many installations that translates into, but it sounds like a lot. If you look at HCI and what's happening with vSAN I think it's a success story not just for VMware but for the technology. You don't often see adoption curves like we've experienced with HCI over the past half decade. 01:13 MK: Yeah. 01:14 SM: So today we have a great guest, Lee Caswell. He's VMware's Vice President of Marketing for its HCI business unit, that's a vSAN. And Lee is the man to talk to about what's happening in the space. So thanks for joining us, Lee. What a great time to be in the HCI business. 01:31 Lew Caswell: Yeah. That's true. And thanks so much for having me. Boy, I love your work and it's always interesting just to share what's going on in the market, some of the market dynamics and also particularly with individual products. So thanks for having me. 01:43 SM: Yeah, great. So HCI if we look at it as I said in the intro has become nearly ubiquitous in the enterprise over the past half decade. It started off 10 years ago without VMware's participation and was kind of serving the ROBO market, and what we're now starting to call Edge. But it really wasn't until VMware entered the market which I... Is this true? It was somewhere around 2014 that you released the first vSAN? 02:07 LC: About five years ago. 02:09 SM: Yeah. About five years ago. And then now you're at 40% market share. That the curve began to steepen and adoption skyrocketed. So what is it about HCI that just seems to make sense in the enterprise that this market continues to grow at the rate that it's growing when we're seeing maybe more traditional server technologies begin to flatten off? 02:30 LC: Yeah. It's really interesting around how new technologies get adopted right in. We've always... At least I've always looked and thought that things that go faster and are less expensive move really fast. Things that are faster but are more expensive take longer and become niche technologies. But when you find something that's actually... And by costing less could be it's easier to operate as well. But when you find something like that, what you find is that word of the mouth is the fastest still way that the technology is spread. And what customers have figured out is this. The economics of servers are undeniable and if you take the software that hyperconverged builds, what we're doing is we're basically taking general purpose servers, high volume servers and then re-purposing them where we're taking the old technologies around like what was direct attached storage, if you remember, and now pulling that across server nodes. 03:25 LC: And what's happened over the last couple years is CPU power has grown, low latency networks have grown, and flash has irrevocably changed the way we build out architectures now. And so for the first time for IT folks, storage latencies are lower than network latencies. It's a really fundamental architectural change that's happened. And so as a result, you can now build the scale-out systems that basically approximate a scale-up system. And so what we're doing now is we're just taking servers and using those to build that out. And that's very common, by the way, is the way that cloud builds out, the public cloud is built on servers, not proprietary hardware boxes. And so these are... That first ROI economics of standard servers and as scale economies are really driving a performance and cost improvement that's fundamental. Makes sense from a growth standpoint? 04:25 SM: Yeah. No makes total sense. 04:27 MK: If I could just jump in with a question real quick on that and say, it seems to me that... And maybe it's the cloud dynamic. Maybe it's a little bit more of some other factors. But this notion of IT just consuming resources, it seems like there's been a shift in the way IT operates. There was a great comfort in buying those racks, buying those storage boxes, networking gear. It was there. That was IT's job and it was their function and life. As they've shifted from this break/fix organization to this enablement organization, their ability to just deploy and consume resources has kind of... It feels like that's fed into this feeling of HCI as well which is you see everything becoming as a service like. Would you... Do you see any of that in your customers, Lee, or? 05:24 LC: Yeah. One of the things that's been interesting is IT as a cost center was... Probably made a lot of sense in the times when applications changed relatively slowly. So when you had ERP systems and payroll systems and things that change every five years, well, you could plan around five-year time horizons. And that made a lot of sense in the time of where applications weren't changing fast. Now what's happening is the fastest element of change is the application pace of new development and deployment. 05:57 MK: Yes. 05:57 LC: And so when you think about that now, you've generally wanna organize around the element that's got the fastest beat rate. And so as we watch the application churn, the number of new applications, every company is becoming a software development and apps company in the sense. If you look at things like the new CEO of Nike came from where? The IT industry, that's fundamentally changing the concept of what does it mean to be successful even in retail, for example, or financial services? And so that gives you an element that you need an... Infrastructure now has to change at the pace of applications. This is one of the areas where HCI or hyperconverged infrastructure, allows you the flexibility to go and scale on server increments. And basically it is designed, and we don't say this publicly probably but it's for people who can't or don't wanna plan. 06:52 LC: When you think about especially in today's environment, think about how people are now changing their plans based on COVID. All of a sudden you have to basically make these radical shifts in how people are working, that's today's environment just captured encapsulated to a T. And so hyperconverge infrastructure gives you that quick twitch response time that allows IT now to be in the driver seat for building revenue, protecting security, driving new application development. 07:20 SM: And that's a value prop I think that's changed over time. Originally HCI was more about the appliance-fication if you will, a certain data center functionality, which is why it found a nice home in ROBO. But as we move more toward hybrid cloud, the kinds of things we're talking about where speed of deployment, software defined everything, really becomes critical. Do you see a bifurcation of HCI between what we're doing with Edge and what we're doing for hybrid cloud, are these two sides of the same coin for, as you look at it? 07:50 LC: Yeah, I think your assessment first was that HCI basically came in outside the data center as the first landing spot. It was remote, offices, management clusters, VDI applications, and those are areas that needed a lot of performance but they're very cost sensitive, and you didn't have a lot of expertise sitting out in let's say hundreds or thousands of retail stores, or on an oil rig. And so those became environments that were really benefited from the cost single SKU scalability of HCI. But then something interesting happened right around how customers started getting familiar with the operating model, that simple operating model. And basically what companies have done is, they have converged their IT organization, and that's what we find as both the biggest benefit, but it's also the biggest obstacle. 08:47 LC: So the obstacle is having three-tier organizations, where there's a storage team and a server team and a compute team, and an application team. What we're finding is the ability to go and converge those teams, is what gives you the ability to respond really quickly. And now, our performance issue is diagnosed from a single management pane instead three different siloed organization. So, that speed of deployment rapidly really important. 09:13 SM: I get it, and the remote manageability. I think, as we're recording this, we're in the midst of most of North America is locked into their homes, working from home. And we're reading stories every day and hearing stories from our customers about, "How do I manage my data center?" Because now suddenly, every enterprise data center is like an Edge deployment, it's that software defined infrastructure as we go in through a vSphere console, and really reconfigure it as I need it, so makes a lot of sense. 09:38 LC: Well I think, we do go through these stages where things centralize, and then they distribute again. And certainly there was an early phase one of the cloud where people thought, "Well this is just all gonna go to a single central place." Maybe like an electrical utility. Now what's happened though is, people have realized that the three main elements we call in the law of economics, which is that, boy, it's always more expensive to rent than to buy if you know what you're buying. There's a law of bandwidth, which means that you need to have local compute, everything can't rely on a cloud hop going into, across the network. And then lastly, is what we call the law of the land, but that's really the idea from a sovereignty and security standpoint. Data people particularly know that your data and the ability to go and own it, control it, know where it's replicated to, these are key elements. 10:31 LC: And so that means for our customers that we see this, what we call the hybrid cloud. That means there's gonna be Edge to core to public cloud, tap into the public cloud expertise, by all means, but you're gonna have this distributed environment. And it gets even more interesting when you get to like an IOT Edge world. 10:52 SM: Right, and directly related to that, I think you announced just last month and are releasing it now but, VMware Cloud Foundation 4, VCF4 is launching, and that's tightly integrated with vSAN. Can you talk about how, or explained to us and our audience, how vSAN and VCF fit together? 11:15 LC: Yeah, let me start with... So vSAN took the same architectural approach that we had with vSphere itself. VSphere if you recall took the dominant position in the hypervisor market, because we had an architecture that basically took the hypervisor and integrated it with the Linux kernel. And similarly with vSAN, we took that integrated approach where it's a hypervisor native approach, not everything is in the kernel literally. But the idea is that by being integrated with the hypervisor, we give you the performant, efficient use of resources in a single source of management. And so now vSAN, which is adding over 100 customers a week, and it's taken the number one position in the market by far, pulling away from all over the competitors right now, basically now has the opportunity to go and extend beyond just storage and compute. 12:07 LC: So if you think of storage and compute, we took the industry leading hypervisor vSphere, we have the industry leading HCI stack with VCM. And then what we looked at was the same dynamics that are leading to serverization of storage is also leading to a software-defined approach to networking. So why is that? Well, it turns out like scale-up networking boxes for load balancing or firewall protection or security, were basically single choke points for performance, basically hairpinning, and lend themselves to the same software approach where we could have a distributive software approach. And so our VMware Cloud Foundation pulls all of the resources across the major fundamental compute storage and networking elements and wraps it together with both management and the ability to life-cycle manage all the pieces together. That's what VMware Cloud Foundation is. And so VMware Cloud Foundation now is a full stack, fully integrated, deployed on servers and allowed basically on-prem or is an identical stack available in the public clouds. And this is giving customers just incredible flexibility to have a software-defined data center, which we've always looked for and now is available today. 13:25 MK: You're talking true definition of hybrid cloud at that point, too. Kind of minimizing the pains of less people, our IT folks have to utilize to manage their environments and the replication of environment, both on-prem and in the public cloud as well. 13:40 LC: Yeah. 13:41 MK: Versus the watered down the definition a lot of folks have out there of hybrid. 13:45 LC: Yeah. When you look at the public cloud and you think, "Well, how do I integrate that without creating yet another silo?" The way to do that is to have the same operational model, we call consistent operations. And so consistent operation means that, "Hey, I can demotion things for example, from on-prem into the cloud." I can do failover where I can go and basically have storage replication, and failover fail back across the hybrid cloud. And now I can do it with the same operational team. Well, some people at times have criticized VMware for lock-in on the hypervisor. 14:20 MK: Yeah. 14:20 LC: What we look at is, we're providing an architectural advantage that is of real value and we're giving customers the degrees of freedom they're most likely to exercise, which is choice of cloud, which is choice of applications with our newest Kubernetes integrated capabilities. So now you've got those degrees of freedom and that's what's really driving the market leading adoption right now as customers are feeling that level of operational leverage in a time where they're trying to do more with the same or even less people. 14:52 SM: There's a couple of points there I wanna hit on. But you said they were an operational lot. And we looked at, as we watched the launch announcement for vSAN 7 and VCF, you used a phrase that we don't see a lot of vendors use, which is I'm gonna try and quote this right, "consistent reliable day two operations". 15:11 LC: Yeah. 15:13 SM: What does that mean? What does that mean to you? 15:14 LC: Well, there's the first... First element is "How do I go and actually install and deploy things?" And particularly, in today's environment, for example, how do you minimize the number of people it takes to go and put something together? A common integrated stack means that, "Hey, I've got the integrated capability." And we're working with our hardware partners. We have a big, broad ecosystem and of course we work very closely with Dell EMC with our VxRail product, where that actually integrates the full stack together. So that's day zero. Now day two means... Alright. Well now, let's say six months down the road, one of the elements of our software stack is upgraded. I've got a new vSphere release or there's a new vSAN release or there's a new vRealize release as part of our management stack. So rather than having to go and basically update each one of those individually, that day two operation is really simplified when you have a life cycle management component. 16:13 LC: We call that SDDC manager, and that's the core IP involved in VCF or VMware Cloud Foundation and that's what allows you to go and get this one click upgrade across all the components to help with that day two operation. So people can look... Stop looking down at the infrastructure and look back up to their apps. 16:31 SM: That makes a lot of sense. That makes a lot of sense. The other point you hit on that I wanna touch on, you mentioned Kubernetes. And one of the big key features of VCF4 is the integration of what you used to call project Tanzu into your product line which brings some sort of container management, some level of container management into VCF. So how or do vSAN and containers play together? Is their functionality beyond just delivering some CSI drivers for storage and calling it a day or is there deeper levels of integration? How do you think about vSAN and containers? 17:06 LC: Yeah. A lot of the early thoughts on containers were that containers are gonna wipe out VMware, just like the cloud was gonna wipe out VMware. The idea was that, "Hey, here's another way to basically think about a pseudo-virtualization, if you will." Here's what we found in practice. Containers are incredibly valuable for developers in speeding and making development more flexible. And so most new applications, in fact, some data shows 95% of new applications are built with containers in order to speed development times. Now at the same time, the speeding of development times was actually complimentary and entirely aligned with the speed that hyperconverged infrastructure offers in terms of making a flexible, easy to manage infrastructure underneath that container development. 17:57 LC: And so, you've got these two complementary elements. And so what VMware has done is we've taken our vSphere flagship products and we've redesigned that, re-architected that where Kubernetes is integrated into vSphere. That's a fundamental architectural enhancement that allows us now to go and offer the same policy-based management for containers that we offer with VS. And so that element is first. And then with our Tanzu Kubernetes grid and our Tanzu runtime services, those are the abilities now to go and start looking at how to present Kubernetes APIs externally to developers. And the idea overall is this while developers may not be our ultimate customer, developers are coming to IT and asking to be responsive to these new applications. 18:49 LC: And what we're offering is that those developers can now come into a VMware Cloud Foundation environment and be given an secure with intrinsic security environment that allows you to go be as responsive to container-based applications as we've always been for traditional VM-based applications. 19:09 MK: Yeah. That makes a lot of sense. And we look at containers... And you're right, there are a lot of folks who look at containers as, "This is a VMware killer." And I think one of Dell's big competitors of today. And this is the answer to the VMware text but the reality is they serve very different functions. Containers give me the ability to quickly package and sandbox an application. And coming out the DevOps world, it is just largely stateless becoming more stateful. But that's entirely different value prop than I get from a hypervisor which gives me the flexibility to build an infrastructure. So, to me, the vSphere in the core VMware gives me a software-defined infrastructure, whereas the containers give me a little more application flexibility. 19:52 LC: That's a good insight, Steve, where one of the things we're finding is that new cloud-native developers are working with containers actually have an important and sometimes final say on whether the applications they develop will eventually run. So they look on-prem and say, "Well, gosh, on-prem is too hard." Containers have some pretty gnarly characteristics. They're spun up more often, spun down, there's no inherent security bringing them together. And so if VMware can come back and say, "Listen, you can develop where you like. You can develop in the public cloud, you can deploy responsibly on-prem, you can basically have a bridged environment." Well, now I've got a really interesting way to go and make sure that those new developers are able to go and deploy in a hybrid world and take advantage of, "Where does it make most sense either from economics or bandwidth, or even security purposes to [20:46] ____?" 20:46 SM: Or even ease of deployment and trainability of your workforce. If you look at who manages IT infrastructure, there's tens of thousands of people, IT administrators, who know how to manage an infrastructure using vSphere. And those guys don't wanna go learn another tool. And there's risk in having them do that, as certainly as we go from zero to proficiency. I love what you're doing with the integration here. I was a fan of this from the time you announced it at VMware of last year. 21:14 LC: One of the things I love about VMware is we are leading into things that were perceived as headwinds and looking at how to make them as tailwinds, so the cloud. People were like, "Well, wow, cloud." Well, we came back and said, "Listen, we're gonna help our customers even if they wanna migrate everything from on-prem to the cloud. We're gonna help them do that. We're gonna help customers find the path knowing that, in our view, the hybrid cloud is the answer because of these distributed realities." 21:44 LC: And then, one last thing is that a lot of our customers are, frankly, wrestling with how to go and attract talent. So if you can basically bring the talent that you have today, extend the capabilities. I also learned early on if you can make your customers... Give them a successful career path, you got a long history in front of you. 22:06 SM: Yeah. And we're guilty. Matt and I are both industry analysts and I remember when you started moving toward the cloud and even containers when you talked about Tanzu. Originally, it's like, "What the hell are these guys doing?" But then you think about it and you break it down to constituent parts, and it really is all about software-defined everything. And you have all the pieces with... And you talked about networking with NSX earlier, and all of that comes together to make sense in terms of how I deploy infrastructure. Go ahead, Matt. 22:35 MK: So, speaking of software-defined and infrastructure in general, going off on a tangent a little bit but I know I've talked to a number of folks in IT and there's a bit of... Maybe confusion is the wrong word but there's a notion of... There's HCI and there's CI. And it feels to me like as VMware with VxRail, with vSAN, with VxRail continues to expand its capabilities, there's this demarcation between CI and HCI that maybe seems to be going away a little bit and I think there's a lot of confusion. I'm in an enterprise. I'm evaluating, say, VxRail. A Dell rep comes in and says, "Really, given your size, you really should be looking at PowerOne." Or an HPE rep comes and says, "What about Synergy? You can run VCF on it." Can you from a VMware perspective, is there a line between the two or are there use cases? Or not use cases but... 23:40 SM: Demarcation? 23:42 LC: Yeah, great question. So the CI market which was really pioneered by Vblocks, FlexPod. And it was basically like, "How do I take these pieces and hold them together?" But at the same time, a converged infrastructure product still requires the independent management of storage. And in fact, you can use tell because, normally, you have to license by terabyte and manage that storage by LUN file or object. That's the definition of a storage system. And so you're still managing, you're still planning around storage. And these pieces although there are orchestrators and things to make it easier to go and deploy, there's still individual siloed elements that are pulled together. 24:28 LC: HCI, which actually now has just passed the CI market or passed all the other converged products right now in terms of market size, has a fundamental different architecture in that the elements are inherently together. So when you're scaling out a system on HCI, you're scaling out storage and compute. And there's certainly levels of flexibility. You can add more compute and more storage based on the nodes you select or based on supporting for external either initiators or external VMs. But in general, you are scaling up storage and compute as you grow. And so, that allows you to have an integrated management architecture that now gives you a single management point. 25:17 LC: And that's distinctly, concretely different. Where if I'm looking at a VM in perspective, I do that from a HCI console. I start from a VM or now a container. And I can go and look at the attributes. It started off with CPU and IOPs, or CPU and RAM, and then extended with storage into IOPS and QoS capacity. And now with networking, it's extended into networking capabilities and security that move with VMs. That's very different than managing a LUN separate from a storage... From a VM server from a different network stack. So I think what we're seeing... This is why HCI is following such a rapid adoption pattern is the simplicity of use has really caught users' attention. 26:09 MK: So here's a question, and I'm trying to phrase it without sounding like a big VMware fan, or otherwise. But there's been a time in the HCI market, where there were other vendors that had a lion share of the enthusiasm, if you will, market enthusiasm. 26:31 LC: Oh, sure. 26:31 MK: And I'm not gonna say quietly, but you wake up one day, and lo and behold vSAN is killing everybody. [chuckle] 26:37 MK: What do you tribute to that? Is it kind of the experience that VMware brings from a manageability perspective, and just overall, your roots within the enterprise, or... 26:49 LC: Yeah. Good point. Alright. So VMware wasn't first in the HCI space. And in the early days of the HCI space, we basically had a set of early moving competitors who operated architecturally by the way, if you went back to my earlier statement really as what we call a type-2 HCI, which is basically a storage stack that runs on top of a guest OS. Similar to how, if you go back and have been around vSphere a long time, GSX, for example, Workstation, were type-2 hypervisor instances. And that integration, which took some time to go and basically get to be enterprise ready. We released that really in a 6.2 time frame roughly four years ago. At that time, we added a whole set of storage services, and have continued that rapid torrid pace of development. Simple things like encryption, compression, deduplication, different FTT levels for failure protection. We just introduced with vSAN 7, scale-up file services. So you now have a file, a name space that crosses node boundaries. 28:03 LC: These are all those types of services that we've captured over time. And so what has fundamentally shifted though, and right now why VMware and vSAN have more customers than the next three competitors combined is the fact that HCI is being adopted as enterprise infrastructure. 28:25 MK: Yeah. 28:25 LC: When these were just like boxes sitting out on a ROBO area or something, that's mission critical on its own nature. But that wasn't the same as having HCI as your fundamental transactional database-driven infrastructure. And so now all of the support tools that VMware offers, all the enterprise resiliency... And one last thing. In the early days of vSAN, it was just vSAN. But now with our Virtual Volumes or vVols capability, VMware is unique in offering the same policy-based management across HCI and storage systems. And so with our storage ecosystem partners, you can now have a management construct that allows you to go and allocate capacity IOPS, QoS by... From a V-centered console. 29:17 LC: And so we're helping customers do what they really want, which is they rarely throw things away. And so, "Listen, I'm not tossing away all my storage, I'm not tossing away my fiber channel in restructure. Help me go and manage storage alongside HCI and basically make this a seamless transition." VCF, for example, supports both vSAN and [29:39] ____ storage. These are some of the ways that we're helping customers deal with complexity, get more operational leverage. 29:48 SM: So you mentioned the word partner., I wanna switch gears a little bit beyond product offerings and raise a question that we've heard in the market, and increasingly more. Dell and VMware are not shy about touting a better-together story, and this has become more true as time has marched forward. So as IT buyers look to deploy vSAN solutions on non-Dell hardware, and there are still a few guys who like to buy other people's servers and storage, is there any reason for them to worry about being underserved by VMware, because of the hardware they're choosing to buy? 30:21 LC: Well, yeah. It's a great question, 'cause certainly we do a lot of work with Dell. They're a fantastic partner with us. On the vSAN side we have the VxRail product, which is a jointly engineered product, between the company and is doing extremely well. Number one, in the share... Actually it's three times larger than the next competitor in that space. So they're doing a fantastic job of leveraging, hardware integration, support, expertise on that same point. At the same time though, our vSAN Ready Node business and the software running on servers is deployed... It's still the dominant deployment model for vSAN. And that gives customers a lot of choice. By the time a customer comes to us, they've generally decided on a hardware company. 31:07 MK: Yeah. 31:07 LC: And we're describing our software value. And yes, it runs on the hardware of choice. And every partner... And we work with all of the major server companies. We got 15 different server partners... And including HPE, Lenovo, Cisco, and others. All of them work with us because they know that we have joint mutual customer interest at hand. Customers who wanna have a partner's hardware and all of the tools that they've bought into there along with service and support contracts, and at the same time, wanna have a relationship with VMware. And so that server ecosystem, partner ecosystem, it doesn't stop at servers, by the way. So it also works, as I mentioned, with storage companies, it works with data protection, backup, and DR companies. And importantly now, it works with all of the hyperscalers. So while we started with actually IBM Cloud, if you look carefully. IBM Cloud and then added AWS, and then Azure, and then support with Oracle Cloud and Ali, and others. We've got basically now all of the major hyperscalers supporting our joint products together. So that partner ecosystem is an incredibly powerful source of value for customers today. 32:30 SM: Makes sense to me. I just have one... Go ahead. 32:35 MK: On that topic before we move on, Steve. So as I hear you speaking, Lee, what I'm taking away is, if I'm an HPE shop and I'm coming to VMware saying, "Help me deploy HCI." You're not going and saying, "We know you have standards in place, but let's get these qualified on a different box." You're saying, "HPE, no problem. Here's our pair of configurations for HPE etcetera, etcetera." 33:03 LC: Yeah. It's a little known fact. But if you look at where vSphere resides, certainly HPE is a major, major deployment platform. And what we track largely is server share in terms of new deployments. And so yeah, there's a lot of variation in server types and server expertise. One of the things that's interesting, though, is if you love new hardware and I love new flash hardware, new CPU hardware, it's coming to servers first. And it's coming to servers first because I think based on my calculations, there's roughly about 38,000 servers a day. I think Dell makes a server every 15 seconds. So if you start looking at where do all the new hardware technology show up first? Well, they show up on servers. So you get to look at these servers, and server hardware changes roughly every nine months. So as you look at that, you basically get to choose from the best. 34:07 LC: And with VMware, you get the choice of hardware going forward. And knowing that we do integrate closely with our partners and with Dell, but at the same time, customers really value that choice going forward and know that we have a mutual interest to make sure that customers are serviced and satisfied. 34:26 MK: Yeah. And I think it's smart to what you're saying. There are large investments made by these organizations and the tool chains that they use to manage their infrastructure and to try and force somebody's hand doesn't make any sense. 34:39 LC: Yeah. I think right now, if you think about what customers are trying to manage, especially if we have a downturn economy at all. Customers are looking and saying, "Listen, I wanna make sure that I'm picking the best of the best. That I've got suppliers that are gonna be able to weather a downturn." This is not a great time to be in the startup world. If you think about non-profitable, cash-strapped startups, this is gonna be a difficult time. And so, companies are really looking carefully and saying, "Hey, VMware, I'm gotta bet on you as being able to support this full set of capabilities and be future-proof for whatever I need going forward." 35:23 MK: Yeah. Let me ask you one more question. Sorry, Steve. Sorry. But you mentioned something earlier, and it just struck me as we were talking about utilizing that infrastructure to its fullest capability. You mentioned ROI earlier. And I think about ROI, TCO. And even just a couple of years ago, there are some negative connotations when you start, not negative, but people go, "TCO, yeah." Everybody's got a TCO story. Is there a... Have you found that customers over the last few years have started to really embrace it? Because there is a huge cost savings you can achieve as an IT org by shifting your consumption model. Have you found that that's becoming more real within organizations, or is it still just a... Something that's viewed by IT as a vendor's trick to try and get them to buy their product? 36:15 LC: Yeah. That's a great question because any marketing TCO and ROI numbers are usually viewed with a skepticism or healthy skepticism, let's say. I do think though that customers do get the economic value of a server-based underlying hardware element as opposed to proprietary hardware. 36:38 MK: Yeah. 36:39 LC: We've gone through that... It's a little bit... 36:41 MK: Excuse me. But inherent obvious savings, and you don't even have to necessarily quantify. You can just get the... 36:50 LC: You grok that. I don't know if you remember... So if you think about how x86 servers took over from Unix servers. And by the way Unix servers didn't go away. There's still Unix servers around. But at the same time, all the volume shifted to x86. And that was predicated on a very simple concept, which is it disaggregated the hardware value from the software OS value. That was the first time. I remember when we were back in the [37:19] ____, and those who were integrating values that separated this. Similarly, for the very first time, HCI has split apart the software value of storage from the hardware value. We've just aggregated those two. And VMware is clearly on the software side. And so, what's happened now is it's allowed customers to go and tap into the commodity economics of underlying servers, that's what gets us in the door. 37:47 LC: But I'll tell you what. It's the operational speed and flexibility which is what means... That's why customers, 40% of our customers buy again in six months. And when they buy the second time, they buy 60% more of what they bought the first time. That land and expand model is an unbelievable testament to the operational flexibility and leverage that customer are seeing. 38:10 MK: Yeah, yeah. [overlapping conversation] 38:13 SM: No, no, no. So, we're running short on time. So just wanna wrap up, maybe, with a forward looking question. Alright, let's look beyond 2020. This is an impossible year to talk about, I think, and it seems like every time we look at vSAN and the announcements you make, where else could we go? But you tell me, where are we taking vSAN long-term? If we look out two, five year horizon, what's your vision? 38:38 LC: Yeah, it's a great question. Well, so first, we're bringing this operational leverage to the mainstream of the vSphere base. I think we've got well over 20,000 customers today. There's well over 400,000 of vSphere customers. So we have a long way to go just to get our customers, today, fully realizing the operational benefits of vSAN by itself. And secondly, next up, right, is the fact that if you like vSAN, you're gonna love the VR Cloud Foundation. And the idea that I can take this now and basically have a fully software-defined infrastructure, means you get a different new operational model that's really gonna revolutionize your on-prem world. And then, the next stage is this, how do we integrate across the Edge and into the public cloud? 39:31 LC: So once you have a common architecture that persists across that, we offer that, that's a check from us on VMware Cloud Foundation. The next piece is, well, what are the differences about what's running in the public cloud or in the Edge? So if you look at the public cloud, we look and say, "Hey cloud native apps," well that's different. So as soon as you provide that same flexibility and functionality capabilities on-prem, you now have the integration, basically, the barriers open up, the water starts flowing across the dam. This is what opens up, from my perspective, the, basically, hyperscaler to on-prem world and then the next piece there is probably going to the Edge. And the Edge, I think, we have complete untapped opportunity around now doing processing for 20 billion sensors coming online over the next year. And so, I think we're gonna have this massive expansion of opportunity where a common full-stack provides the integrated leverage that you need as a customer. 40:33 SM: That's a great answer and I think you're exactly right about Edge. And Edge is a world where, I think, the intrinsic benefits of HCI are gonna make a tremendous difference. So, Matt, do you have anything to wrap-up with Lee with? 40:47 MK: No, thank you. It's great to have on with us. I know the three of us have spoken a few times in the past. It's great to get these conversations out there for the folks in audience, all two of them, that have a genuine interest and real questions around this. So thank you. It was very informative. 41:08 LC: Well, it's great to talk to folks who have such a great appreciation for both opportunity, but also of how do customers get started. So that's what I'll leave you with is, find an area. Every enterprise of scale, we think, has an opportunity to start with HCI, get familiar with the operating model, and then see how you can go and leverage this for your best benefit going forward. 41:28 SM: Makes tremendous sense. Hey, Lee, thanks again for taking the time to talk to us today. I know I learned a lot. I learned a lot. 41:35 LC: Matt and Steve, appreciate it, Thank you so much. [music]