 Live from San Francisco, celebrating 10 years of high tech coverage, it's theCUBE. Covering VMworld 2019, brought to you by VMware and it's ecosystem partners. Hello everyone, welcome back to theCUBE's live coverage here in San Francisco, California for VMworld 2019. I'm John Furrier with Dave Vellante. Dave, this is our 10th year, 10 years covering VMworld, quite a run. Got a great story, more stories coming. VM&As, a lot of organic growth, a lot of IPOs on the startup scene. Our next two guests, Mike Adams, senior director, VMware and Ziv Kalmanovic, product line manager of VMware. Welcome to theCUBE, great to see you. Yes, great to see you too. You guys got a lot of activity happening around Bitfusion, a lot of news to share. Exciting, the M&A story has been high on VMware. We talked with Pat Gelsinger earlier. Yeah. Continuing to fill in on the strategy. Yeah, absolutely. Give us the update. Yeah, I think the key thing for us is we really want to become a key player in the AIML space and say that those workloads should come on vSphere. And with this acquisition, we think it provides a great framework for a lot of the hardware accelerator devices. The best of use are known of those as GPUs, but we think there's forthcoming market with FPGAs and also custom ASICs. So we're super excited about that. For the folks that don't know much about the acquisition, what was the motivation? What was the company's core product? What was the interest? Yeah, the company had a product called FlexDirect and that particular product was really focused on taking a similar concept that a lot of VMwareites know, which was, hey, in the compute space, we were trying to take these isolated islands and pull them together. Same type of thing here. You had these expensive devices that people were buying and they were isolated. And now if we could take a single server, it's got a bunch of GPUs on it. Why don't we share it, right? You see all these papers that come out around machine learning and at the very end, it says, geez, I'm amazed that these GPUs are so underutilized, even when we're actually using them. It's kind of like buying a car and then using the radio only, right? It just doesn't make sense. You've got this trend of alternative processors just sort of exploding all over the place. I mean, obviously in video, it's just sort of people know what's going on there, but you've got ARM, now you've got the edge coming in. Intel's still dominant in the server space, but even storage devices today use different type, not Intel processors in there. It's a combination of ARM or sometimes GPUs, as you say, FPGAs, even though they're sort of a narrow use case. You're seeing A6 make a comeback, so you've got all this additional processing power. So that's a tailwind for you guys and it's sort of the intersection of those two. Maybe talk about some of the trends you see in that regard and how you're taking advantage of them. Yeah, I mean, it reminds me of many moons ago when we had new chips that were coming out and we said, geez, there's a hardware flurry here, right? And now we're at a really similar spot. Ziv and I see a lot of different types of devices and acceleration devices, whether it's compute or network or storage. And in this particular case, right, we just see a hotbed of all these customers that are seeing the same problem, right? And we've got great partnerships with Intel that you mentioned, NVIDIA and many others. And we just want to really leverage those for these devices because you look at vSphere and say, okay, traditional workloads, we've done those very, very well. But as we get into containers, Kubernetes, machine learning and AI, we want these newer cloud native and newer workloads to come our way and taking advantage of these new capabilities really helps accelerate that in a big way. Could you explain more on the vSphere impact? Because, you know, first of all, VMware community, you get the feedback right away on Twitter on a lot of things. But sometimes you got to dig in and find out what people are thinking and where there might think there could be future opportunities or, there could be some skepticism as well. The vSphere native, having AI on vSphere, it's just mind blowing to me. But I mean, I can see a data processor kind of vibe going on here where data needs to be processed. That seems to be a trend. What is going on with vSphere with this? Is there, what's your customers know? Well, I think that the first thing to clarify here is that, you know, often there is this question, why would I run ML or AI workloads specifically on vSphere as a platform? But then customers do run ML, AI workloads on public clouds and those layers are not that different than vSphere. It's virtualization layer and they are running it in virtual machines. So the whole idea with beat fusion specifically is that actually we can make it even more efficient to run these workloads on top of vSphere because the underlying infrastructure that you actually have to accelerate these workloads, today they are mostly GPUs, obviously, but in the future, as Michael also mentioned, new ASICs are coming in and FPGAs are coming in. We are going to make those as well. That's the plan, using the beat fusion framework, be more efficient to use and utilize. I'm surprised that people are skeptical around running machine learning on vSphere, not skeptical because, I mean, it's great for any time you have the opportunity to automate something or use software to make something go away that's not either differentiated or undifferentiated. I mean, so it makes sense, but I'm just trying to figure out where specifically within vSphere is it being targeted to use? Where in vSphere, as in, well. If I'm operating vSphere, I'm an operator. And I got devs kicking around the corner, I got cloud, my whole hybrid clumbing. Where does this fit in? This fits into essentially any place where vSphere is running. It doesn't matter if it would run on VMware cloud or any other for cloud partnerships or at the edge where vSphere runs. This is a core capability of vSphere. So it doesn't matter where physically or infrastructure is, we would be able to expose this technology. The idea is also that you mentioned the trends in the ASICs as they're coming into the enterprise. There's an architectural change that's also coming in in the server perspective. It's just, the servers are actually getting more denser in the accelerator infrastructure that they have in them. So you're seeing four to eight GPUs in a single server. Those are very powerful machines. You can just move oil workers into a single machine. Again, that brings us back to the fusion and this desegregated model of accelerator use, which is very similar, by the way, to centralized storage use, for example. I think you guys are on something really big here. I think that hardware assists offload. Anything hardware assists on hardware offload is going to be a bigger trend. I mean, we saw it happen big time in hyper-converged just for storage and networking. But I think as you move up the stack where Kubernetes is going to flourish. I mean, imagine all the services that have been turned on, turned off. I mean, that's not, I mean, you don't even know when it gets turned on or off. I mean, offload for a while with things like arrays, right? Trying to push processing off to a big array that you've got there. And then one other thing you said that I think was really important is the audience, right? If you look at AI and ML, we traditionally haven't talked to the data scientists, to the machine learning folks. And we need to get to the IT folks that are supporting those workloads saying, similar to some other workloads that were new and saying, these are going to come your way. And so we need to be prepared and you need to be able to leverage them. So what's the pitch to those folks? What's the, what are you guys saying to them? Because it is a benefit for devs and because dev ops is to have an ops, right? And we've got the ops down. Okay, see that and there's some chains happening. But a dev, what's the pitch? What's, how do you get their attention? What's the value proposition? Yeah, you would take that. The, actually, that's the beauty of it. It's exactly the same valid proposition that vSphere and VMware, the VMware stack provides to the developers. The only thing is that now we are letting the, the ops people to actually provide this new infrastructure as well in the same efficient manner. So it's your transfer. Basically it's giving the exact same value proposition. Talk about the, the multi-cloud tie in here. We've heard a lot about multi-cloud. I think multi-cloud in part anyway is being able to run any application and workload anywhere. And one of the things about your technology is the ability to not have to rewrite the application to take advantage of acceleration. Does it fit into multi-cloud? And if so, how? Yeah, when we made the Bitfusion acquisition, if you look at their story, they had the any, any, any story as well, just like we do. And so, you know, we made announcement this week with NVIDIA and AWS and VMware. It's definitely possible with the technology that we have, you know, to extend that even further. And so, you know, the only thing I know with users going forward is they're going to have more than one cloud. And so we just need to prepare for that and make sure that it works and it works well across the board. And the common layer when you look at our multi-cloud strategy is vSphere is going to be at each of those layers. So if it ties into vSphere, it should be pretty easy to make it work in each of those environments. What was the, what was the announcement you made? Can you just share that with us? Well, the big one was being able to use NVIDIA in the context of cloud and AWS. So those GPU capabilities and bring it to the service as we do on-prem. And so that was a big piece. And then we also, obviously, in and making that announcement, talking about, hey, you know, this is a critical area for us because not only are we doing this, but we're also saying that, you know, Bitfusion will help enhance this because we think NVIDIA and Bitfusion work very well together as well. And is that a product, a service, a go-to-market initiative? In the case of the cloud and AWS, it would be offered as part of the service. So when you can consume the compute, you know, you want a GPU, it'll be there for you to help run that workload in the cloud. And that's available when? Well, that's an NVIDIA and AWS kind of question when they are making that infrastructure available. It's essentially going to be another instant style that VMware cloud and AWS will offer. Right, okay. Yeah. So that's a tech preview. Yeah, right. What are some of the things that people should know about? Because, again, in the pattern I'm seeing here, VMworld is, as they move up the stack with Kubernetes being that abstraction layer that Gelsinger is promoting heavily. And rightfully so, we're big fans Kubernetes were there from the beginning. Sure. Is that you're going to have this purpose-built native capability. So you guys have got this native vibe going on. Native to high vSphere. Native to ESX. Yeah. Native, I mean, what does that actually mean? Native, like Kubernetes is native on high, what does that native mean? Explain to the audience what that actually means. Yeah, I'll start up. Sure, you can elaborate. Yeah, I can talk about it for another 30 minutes if you want, but it left like for a second. I mean, what does that mean? True native, native. The idea for us was, you know, use Kubernetes in really two ways. You know, most of the time when we've been talking about Kubernetes in containers, it's running that on top of vSphere, right? Well, what happens if you could take the DNA of that and put it actually inside of vSphere, right? So not only you could run these clusters and native pods, but you could also leverage some of the value. And one of the things that Kubernetes does really well is it handles workloads really well. So if we take an example where we have 145 VMs and they make up your app, right? Normally you'd have to go to each one of those and figure out, okay, let's make some changes and some tweaks. And now what I can do is I can treat all of those as one workload. And I can move them, I can do really interesting things with that. And that's the power, one of the powers that you have with Kubernetes. And that's where the differentiation is. Anything to add there, Zip? Yeah, exactly. I mean, you are essentially getting there are a lot of benefits or customers, the value that the customer is getting today from vSphere, generically speaking. And our long-time customers are familiar with those value propositions. And what we are saying is that when you are getting something as a native capability is that essentially ties into all the other capabilities that you already know very well and you will be able to get those with on top, sometimes on top or with conjunction with the technologies. So what is that going to enable now? Let's talk about the enablement piece. So let's go back all the way, if you go all the way back to Bitfusion, for example. If you enable it as a native technology, then if you are running containers or VMs on vSphere, natively, they can consume the Bitfusion technology. If you have Kubernetes, it can orchestrate natively the VMs and containers that are using the Bitfusion technology and et cetera. So this is the whole thing tied together. It's more efficient from a platform standpoint. And it's easier to manage as well because you don't have to install a bunch of stuff one on top of each other because it's native. It's part of the infrastructure. So a lot of hassles go away that people might know of. That's right, bake it in. And you're going to have two guests tomorrow that are going to go deep into it with you. So we're excited. So we're hearing a lot, obviously, about Kubernetes at this event. But most of the audience are not developers. So how can you use the sort of Bitfusion mojo to attract developers for some of these new workloads that are coming to the marketplace? Yeah, I mean, it's all about acquiring new audiences. In our case of Bitfusion, it's more the data scientists. In the case of the Kubernetes, it's more around the developer. But I think let's use the Kubernetes example as a good one and what we announced with Project Pacific. Basically, the way it looks, the technology looks to them, it'll look like the Kubernetes API with a little bit of vSphere goodness. And from the operator perspective, the people that we know, the 20,000 that are here, it looks to them like vSphere with some Kubernetes goodness. So that's the right mix as you've got to get it so it looks exactly, smells and feels just like what they're used to. And I think that that's a key aspect. And then for the data scientists with Bitfusion, we really need to say, okay, you want to run these workloads, but geez, you're paying really a lot of money for these expensive, isolated devices. And you could get more value by kind of grouping them up and making sure that they're used kind of in aggregate, right? So there's more leverage on the data science side. So if I'm, say, hiring someone, I don't know. Oh, I got it. I'm more to work with. With you guys. Exactly. Essentially, it's the same story. They don't need to change their applications, their frameworks, their models. They use the same CUDA interface, which is the GPU interface for the GPU compute. So let's talk about that. So data scientists, they always complain that most of their time is spent wrangling data. That's their bugaboo. And then there's a collaboration between data scientists and developers which probably doesn't happen enough. What are you seeing in terms of the trends and from the data science role and how can you help solve some of those problems? Well, what we are out to solve is really access to infrastructure for them. Easy access to their infrastructure and their software stack. And the way to get there is to make the data engineers that serve these data scientists and the application administrators that serve these data scientists to get easy access to the infrastructure they need to provide the software. And that's where vSphere eventually comes in. So it's not necessarily a direct relationship with the end users. It's more enabling the entire organization that actually serves these end users and let them use as much infrastructure as they can as possible. And your partners in that end user organization. Yes. That's the buffer, right? Guys, last question. Share what's the plans are. What's next? What are your goals for the next six to 12 months? I'll see. You get the acquisition under your belt, native in vSphere, a lot of other cool things. I mean, obviously, yeah. I can talk about customers and maybe you can talk about product from a customer perspective. You know, we want to engage in proof of concepts. So we want to bring them in, let them test out the software. It already works with vSphere. So I'll be running with multiple proof of concepts across the globe. Any use case or any use case or what's? Yeah, I mean, it's pretty simple at the moment. It seems to be most people that are using GPUs around ML. We have a great demo down on the floor that shows people trying to run in exception three or resident 50. And then how can we actually help those VMs that are running that? So that's going to be my focus for the next six to 12 months. All right, you heard it. You got some use cases. Come over here and bring them up to Mike. Obscuse. And from the product perspective, I mean, obviously we acquired Bitfusion in an early stage. The technology works well. It works well enough to be productized. However, VMware and vSphere has very high enterprise software standards in terms of security and management and governance. All these capabilities. So that's going to be our focus for the next, you know, even almost a year to make sure that we bring it up to a level where we can confidently provide it and sell it as a product to our customers. Yeah, you got an engineering high bar there. Yeah, absolutely. Awesome. Guys, thanks so much for coming on theCUBE. Appreciate the update. VMworld coverage, breaking it down 2019. This is theCUBE. I'm Jaffer Dave Vellante. Thanks for watching. We'll be back with more after this short break.