 from the Silicon Valley Media Office in Boston, Massachusetts. It's theCUBE. Now, here's your host, Dave Vellante. Hi everybody, this is Dave Vellante. Welcome to SiliconANGLE Media's special presentations. This is CUBE Conversations, VMworld preview. We're here at the East Coast offices of SiliconANGLE Media and Wikibon. And I'm here with Randy Arsena, who's the chief marketing officer at Infinidat, and we're joined by Steve Keniston, the storage alchemist. He runs vice president of product marketing at Infinidat Jets. Welcome, good to see you again. Thanks for having us, Dave. So the sands of the VMware ecosystem are shifting, right? It used to be the narrative when Todd Nielsen was president that for every dollar spent on the VMware license, 13 and then 15 and then $18 is spent in the ecosystem. You don't hear so much of that anymore. Things have changed. Pat Gelsinger has now become president, taken the gloves off. They're going harder after the storage. So lots going on. So we're here to talk about what's going on in VMware, the preview of the show. Randy, let me start with you. You've been an observer of this ecosystem for a long time. What are you expecting at VMware? What's new? What's changing? Well, I think a couple of things interesting this year is that obviously with the Dell EMC thing happening, that's kind of the overarching consideration and concern for everybody in our industry, right? We're all watching it very closely as we have been for the foreseeable past, right? So as that starts to come together and finally ends to some degree and we start to move into the next phase of it, I think as a storage provider, and I'm sure other storage providers are looking at it as an opportunity, right? Because there's a lot of uncertainty that that introduces so to your point, VMware had been very aggressively pushing into the storage stack and trying to go after the smaller players and trying to really vertically integrate that whole stack and doing a pretty effective job of it by and large. But this is opening up an opportunity now for companies like Infinitad and others to have a more differentiated solution. And to some extent capitalize on some of that uncertainty because a lot of the customers we talk to are still not quite sure how this is going to execute and what this is going to look like a year from now. So they want to hedge their bets a little bit, right? They want to understand that they can put in a solution today that can live beyond whatever transpires over the next 12 to 18 months and they're not going to be dealing with that uncertainty. They've got a little bit more predictability in the equation. So we think it's an opportunity for independent storage players as opposed to a risk. I think the risk is actually being diminished a little bit by this. Oh, yeah, I was just going to add to that. I think that you make a good point about the $12, $13, $18 being spent in the ecosystem. I think people over time have started to look at where are they taking, where are those dollars being taken up within that ecosystem? Like is it going to storage? Is it going to applications? Is it going to servers? Where is it going? And I think part of the tide that you see shifting is that folks no longer want to continue to spend that kind of money. You see new technologies like containers and that sort of thing that are costing less. You see licensing for VMware changing and coming down a little bit so that people can afford not to spend all of their IT budget on the hardware, but how do they support their application? And it's kind of great. It's funny, we talk about storage. Part of the whole premise behind Infinite Ad is how do we look at the cost of storage as a part of that whole ecosystem? And how is that evolving and how are we playing into that and trying to help customers not spend their whole budget on just storage but help enable them to be able to do other things? Well, I mean, and there's been a 10 year collapse in infrastructure, hardware and software pricing, certainly from a margin pressure standpoint, and somehow the industry just continues to innovate and evolve. But how does that affect an organization like yours? You've certainly seen it due to open source. You've seen it with cloud. How do you respond to that? How do you continue to add value? Well, I think there's a couple different ways. Well, there's several different ways, but the key things that we constantly look at and from my perspective, where we're evolving very rapidly right now is we're kind of shifting away from the importance of media. So the all flash data center, all flash arrays, all flash everything has been kind of the de facto message for the last five years at least. But we're seeing now that even as media continues to evolve and the roadmaps of the providers, the aerial density is increasing, the cost is coming down, although we still have not reached parity and we'll talk more about that, I guess, but so media is becoming less differentiated and more commoditized. So what that does is put more of an emphasis on architecture and software and the services associated with storing and managing data on that media, whatever it happens to be. So you can preach kind of like the importance of flash and all flash as the future, but ultimately the media becomes less relevant if you don't have an architecture that supports the use cases and models of future applications, right? So we think our focus, we believe our focus needs to be on the architecture, on the software, on the scale, on the ability to deliver those services with high performance, high reliability, at scale for the enterprise. So frankly, we are less concerned about the media. The system will work with any media that we want to throw at it. It's currently a dual mode hybrid system that uses both spinning disc and flash. In the future, if the economics of flash makes sense, we can easily adapt it, right? So we have an adaptable architecture and that is where our IP lives. That's where the brain power goes. So, and we'll come back and talk about that some more. So Steve, when I, first VM world show I ever went to, we joke, we should call it storage world, but it was clear that the main problem, a huge problem anyway was storage. The IO Blender, you and I have talked about this a lot on theCUBE and other situations. So storage was very hard. It remains very hard, but as Randy just pointed out, things are changing, less focus on the media, more on architecture and sort of solving customer problems. So first of all, why is storage so hard and what is changing? Part of the reason that makes it really hard is all the different vendors that are in the market who have all of these different types of offerings and they're all preaching, you need this type of technology for this type of application or this type of reliability or whatever. And now it becomes confusing, right? And so I think what you've seen over the last few years is a way to try to simplify that. And a big part of what we were just talking about a little bit earlier on the ecosystem and the difficulties is some of the things that are driving that simplicity are some of the different trends going on in the market today. So take cloud, for example, right? You need to be able to be a lot more flexible and have a lot more simplicity when it comes to cloud. People want agility when it comes to why do I want to deploy a cloud? You talked to CIOs, they want it for agility. Well, that agility needs to both transcend all the way down into the provider so that they have the flexibility to be able to offer those types of services. That turns around and forces the storage providers to provide more and more simplicity into their architectures. So why has it been hard? I think it's been hard because there've been a lot of different technologies out there. People, you see different things being introduced like tearing, tearing becomes important. Compression becomes important. Data duplication becomes important. You see all these storage services start to stack on top of one another. It's difficult for the consumer to understand what is it that I need and I want and it's good for my workload and my data types. And it's difficult for the provider, the vendor, to be able to integrate all that to make it simple for the end user. And another thing is if you look at most storage systems that are still in service, I mean the bulk of the industry is still being served by enterprise storage systems that were architected, not coincidentally by the same people that are at the core of this company's R&D organization while at EMC, 25 years ago. So that fundamental architecture has remained relatively unchanged and is still the predominant architecture in the industry. So what's happened is you've had this massive shift in evolution and convergence of use cases and consumption models and distribution and delivery models constantly being layered onto this legacy architecture and it's just straining under the weight of all this change. So you need or the industry really needs genuine foundational innovation. And if you look at the way we've approached this problem domain, we've looked at it from a very different angle. So we're approaching it at the ground level, blank sheet of paper and developing a system that can evolve into these new use cases and is in fact designed specifically for them. So that underscores the changes that we've seen since let's say late 2000s where the big battleground was who can integrate better with VMware, VAI and VAVP and all these APIs that people would clamor to get a hold of and say we're first, we're faster. Okay, and what happened ended up happening is you were layering on legacy architectures and essentially bolting it in. Hey, it works. Great, and what you're saying and Steve, it plays on what you were suggesting is that cloud now becoming sort of the de facto standard, what people are trying to do is mimic that agility, that flexibility, the simplicity. So you're saying that's kind of the architectural focus and shift, maybe you could add some color to that or elaborate. I think you hear a lot, at least from my experience in the industry or at least over the last two, three, four or five years, you're hearing a lot more about things like REST APIs. You didn't hear about those things five years ago. Yes, every application or provider had APIs and you tried to work to integrate with them and it was always a difficult process. You had the vendor, the two vendors, the application provider, the storage provider, they would work together to be able to provide that type of integration and it was a process, right? Now you're building on top of REST, you're really defining that simplicity. Why? Because everybody wants it and everybody wants, all the hardware vendors want to be able to support it because they want to be the ones to your de facto standard, helping those folks out and they got to get there quickly and they got to get there fast. Well, and from the cloud service provider, any kind of managed service provider's perspective, they need to standardize and automate. I mean, that's their business, right? If they can't deliver on those efficiencies, then they're not going to be competitive, especially when you talk about kind of the mid-tier players. I mean, you got the Apex guys at the top of the food chain that have kind of perfected the model, they've got the land grab, I mean, they're fine and they will continue to grow their business. The cartel. Yeah, the cartel. Although they are increasingly under attack and being chewed at from the bottom by some of the mid-tier. They continue to throw off cash but they're really not growing that much. Right, so. You guys are growing like this, being smaller, great, but. So you look at some of the mid-tier providers and their stock and trade and their kind of, the coin of the realm for them is efficiency and agility to Steve's point. So if you can deliver a system to them, that enables them to subsequently then deliver end user services that are cheaper, faster, easier to provision, easier to manage and more cost effective, then kind of everybody wins. Their end user wins because they're getting a better pricing and easier consumption model, they're getting better service quality, the provider is getting better economics. We as a kind of arms supplier to them are getting a larger opportunity to work with a broader aperture and a bigger addressable market. So it's kind of a, it's a good win for all of us, I guess. What about hyperconverged, you know, hyperconverged, software defined, it's also the buzz. They're clearly trying to mimic cloud. You know, I saw an ad recently, it was like, you want Amazon, you need VMware, you'll love us, you know, where the hyperconverged supplier. Their effort is to reduce complexity and we sort of all know that narrative. What's your take on hyperconverged? You know, where do you guys fit, what's your point of view? I think as with all the different trends that are taking shape in the industry today, hyperconverged, cloud, AFA and all that sort of thing, I think they all have their place, right? And it's really dependent on what your business needs are. Again, this kind of goes back to the complexity conversation of all these different vendors, offering all these different solutions. But if you're a small IT shop and you need the flexibility to be able to grow simply and scale simply and you don't have a lot of resources, manpower to be able to put those things together, hyperconverged is a pretty good solution, right? It's the ability to be able to quickly to bolt these things together and kind of get that scale that you need without doing a lot of manpower and resources. And if you have a pretty well-defined workload or set up a couple of workloads, it's a simple easy solution to deploy. Not every shop is like that. Same with SDS, right? Some folks, some large IT shops have told us, right? We already have the manpower to be able to build our own solutions. Just give me the software. We'll integrate and glue everything together. That's 2% of the industry, right? Not everybody in the industry has 50 development type storage people sitting around that can do that type of an integration, right? So these needs fit different use cases and I think what, you know, part of what we're trying to do is drive that simplicity that focuses on all of the different use cases as best we can. I think that's kind of how we're a little bit different. Well, and hyper-converged largely. The most part is the storage play. Not in all cases, but it's a big storage play and it's, like you said, it has its place. But you can add value, you know, other ways. What about all flash? Let's come back to that. So this has really been a really interesting discussion between the folks at Wikibon and the folks at Infinidet and Moshe came in a couple of months ago. You know, you guys, anyone who follows David Floyer knows, he was first on the whole, you know, flash bandwagon, how it was going to replace high-speed spinning disc. He made that call correctly. He also started to promote the notion of this all-flash data center. Well, before most people, you guys came and said, you know, we have a different point of view and it actually changed our thinking a little bit. He's revised some of his forecasts. Certainly, dramatically increased the length of life for enterprise, spinning disc, directly as a result of our conversations. He went out and did some other research with HGST and WD and Seagate. Your view is different than most. Most people say, all-flash, all-flash, all-flash, go. It's kind of like a brute force approach. Summarize your take. Yeah, so brute force is an apt term, I think, to describe it. So again, I talked a minute ago about the importance of architecture and how the focus is shifting from the media, like to the persistence layer, to the delivery mechanism, right? So it's kind of the network that gets the data where it has to ultimately live. So the architecture that we've developed has been designed from the very beginning to deal with very large, like petabyte scale, very, very large workloads and data sets, but be able to deliver best in class enterprise performance. So a lot of our messaging, and I won't let the cat out of the bag, but you'll see at VMworld, we have some very aggressive messaging around our ability to position relative to all-flash arrays. And again, most all-flash solutions, the example I use is there are a lot of 1972 Volkswagen Beetles with aluminum Ferrari racing engines in them. So you take an old architecture that's 20 years on and showing its age, and you push it as far as you can with mechanical disc, and then you say, well, all we can do now is put faster media behind it. We can't tighten up the architecture anymore, kind of is what it is. So we've got to go to the media to try to speed it up. Now that's okay to a point. And by the way, there are, just as there are use cases for hyperconverge, there are absolutely use cases where an all-flash array is perfect. We go into customer sites all the time, and they'll say, well, look, we have this trading app, or we have this little corner case over here that it just has to live on an all-flash array or a highly optimized purpose-built appliance. And our philosophy is that's awesome. You should run that on what it needs, if that's what it needs. Absolutely. So our philosophy, however, is that typically represents at most 10% of the workload payload in an environment, at most. We're happy to take the other 90. And in fact, in a lot of our customers, that's exactly what we're doing, is we're going in and we're saying, if that's what that needs, although there are lots of situations where we'll take those workloads and put them on Infinibox and demonstrate parity on performance with a higher level of availability and scale. So that aside, if they want to keep those isolated on all-flash, that's totally fine. We are happy to go after the larger population of workloads in most enterprise environments. The other factor is that flash, if you think back five, six, seven, eight years ago, every year, I used to say every year, next year is the year that flash achieves cost parity with spinning disk, it's happening. It's going to be 2006, 2006 is the year. Okay, 2007. Well, it hasn't happened yet. And the reason it hasn't happened yet is that the hard drive manufacturers haven't stopped innovating either. So aerial density, performance, reliability, form factor, power consumption have all improved and are continuing to improve. So while flash is getting better, it's getting denser, it's getting cheaper, it's not getting faster, better, denser, cheaper fast enough to keep pace with spinning disk. And by the way, on top of all that, the volume of data that customers are managing is not slowing down, it's growing geometrically. So when you kind of converge all those things together, you have massive amounts of data, demand for very high performance, you don't have economic parity. The only way all flash solutions can attempt to deliver economic parity is with very aggressive use of data reduction technologies, compression, de-dupe, highly non-deterministic. There's a tax associated with implementing them. If you don't get the compression ratios, they guarantee you have to be chewed up somehow, which is good, the vendors to their credit are stepping up and offering remedies. However, if the remedy is, we have to give you another system, well now you've got two systems. You don't have a federated cluster, you don't have a petabyte class system, you've got a bunch of little systems that are independent that you now have to manage. So it's kind of a solution, but is it really the optimal solution, right? It makes me think of, I mean, with a lot of things, there are two answers to everything, right? So very technical, very intelligent reasons why from a media standpoint, as Randy just pointed out, that it makes perfect sense to do what we're doing. There's also the business side, right? Randy said perfectly, data is growing unbelievably, right? We remember probably in the days when we said a megabyte, a gigabyte, a terabyte, we're never gonna get there. And now we're talking about exabyte scale environments, right? If you're a CIO responsible for the economics of your business, right? And the budgets of your business, you can't achieve financial stability around your storage environment, not chilling through your entire IT budget. If you're not doing the smart things that relate to lower cost media. And we've also all seen the curves, right? The curves say once data has been active or whatever for 24, 48, call it a week, right? It becomes stale. Does that mean I don't need it? No. And in big data environments today, it's like, no, I wanna recall that three month data and I wanna do quick analysis on it. Great, you can't do that from tape. So it needs to be on an expensive, very performant media that enables you to do your business and get those results and those answers. So you guys made a bet. The team that developed the product made a bet, said, okay, we see these trends. We are going to bet that on A, we're gonna make an architecture that can handle any media. But B, we're gonna actually do some things to give us an advantage because we believe that spinning this is gonna remain the cheapest, lowest cost approach. So you made that bet. And that is giving you an advantage in the marketplace right now. What happens if that bet is wrong? What does that mean for you guys? Well, if the bet is wrong, the worst case scenario is that we will, and it's not even the worst case scenario, it is a possible scenario is that the economics will begin to shift to the point where all flash makes sense. If tomorrow, somebody who waved the magic wand and we could deliver a petabyte scale all flash system with economic parity to our current hybrid system, we would do it because our system is 100% software. Every piece of hardware in our system is commodity. There is nothing proprietary. Rip out the drives, replace them with flash media and we're done. The software works, architecture works. Let's flip that. Let's say there are others out there who made the exact opposite bet. What if they're wrong? Well, if they're wrong, the cost of innovation or the innovation hurdle for them is much higher than it is for us. We've done the hard work already. We've already developed this highly adaptive, highly agile, highly modular architecture, very, very good performance, extremely intelligent placement optimization and the things we do within our core Infinibox architecture. So we've done the hard work already. So we sort of front-end loaded the process. When Moshe started Infinidat with this vision of serving petabyte scale workloads with the same performance as a flash-based device, but at seven nines of reliability in a truly enterprise-hardened platform, he got the smartest people he knew together and he knows a lot of really smart people, as you can imagine. He funded the company initially. He sent them off into a nice research facility. He said, take as long as you need. Literally take as long as you need. I'm not looking for a timeframe. I'm not giving you a deadline. You do what you need to do. Here's the problem. Now go solve it. Figure it out. Figure it out. Took them about a year and a half to come up with the core architecture. Took another year or so to get the product to market. A little bit more than that actually. So the approach that was taken was not, all right, look, here's the architecture. Now you go figure out how to fuel and instrument that architecture to make it as quick as possible. And the sort of easy answer is, well, we just put faster media in it, right? The question or the problem Domain has stated was, here's the problem. Forget about what the raw materials look like. Solve the problem in the best possible way and we'll figure that out after, right? So that's the approach. The reason why this is so significant to us is it comes from the person and the team that back when RAID 5 was all the rage, you had to do RAID 5 because that was the least expensive way to do it. They come up with symmetrics and it was RAID 1. And they went, you're mirroring? Yes. And there were really good reasons why. And so they made that bet. Now, past is not prologue, but it caught our attention. So it's a classic Moshe, frankly is really based company way of solving problems that's non-intuitive and non-conventional. Yeah, and I mean, Moshe kind of got where he is by being pretty clever guy, right? I mean, he understands the industry. He is very intimate with customers. I mean, he spends a lot of time still. He spends a lot of time in front of customers. It's where he prefers to be. So he understands that when he's across the table from the CIO of a Fortune 100 bank or health insurance company or retail provider, he understands their problems because he talks to them and he's been doing it for three decades. So his approach is always to understand the problem domain. And he doesn't, there's no sacred cows. He doesn't get stuck in his mind. There's only one way to solve this problem. We can work around the edges but this part has to remain intact. He always starts with a blank sheet of paper. XIV was a blank sheet of paper. Infinitav was a blank sheet of paper. Symetrix was a blank sheet of paper. So he does not allow the past to drag down the innovation process. And there's risk to doing that to your point. It's essentially your role in the dice. It also can slow your time to market because it's fast and easy to get a solution to market if you just work around the edges of the core architecture. When you're starting with a blank sheet of paper and kind of rebuilding the foundation, there's risk. It takes time, it takes money. It slows down the release of your product to market. But if you bet correctly, which Moshe has a history of doing, then what you do is you leapfrog other solutions that are in the industry and it gives you that lead time. It gives you that kind of advantage of being first to market and being able to capture those opportunities. Fascinating chess games. We should do a crowd chat on this topic. We should do a crowd chat on this topic. All right, let's wrap. What's going on at VMworld? Hold on, hold on, wait a minute. VMworld, I hear this little rumor. How many years have you been going to VMworld? Many. And you're not going to share. I think I've been to VMworld more than I have Comdex. And you're not going to share. I can't make it this year. No, I'm doing an event in Boston and so I drew the short straw, a long straw. You know, it's a great event. So it's a big, big, great event. It's not going to be the same. I always said that it used to be Labor Day started the Q4 selling season and now it's, you know, VMworld, right? So we'll miss you, but it's okay. Tell us what, you know, people can expect at the show. So we upped our sponsorship this year. We have a Gold sponsorship. We're at booth 2211. It's going to be a really nice, large booth. We actually have a corner of our booth set aside to be able to have customer presentations who will actually talk about how they've chosen us over all flash arrays and some of the rationale behind it, right? So it's not only just, are we talking about it? Randy said he didn't want to give away the, you know, didn't want to give away the secrets, but, you know, it's been out there on social media. So faster than flash and easy on cash. That's our tag for the show, right? And we're going to prove both with some of our lead technical people, you know, the specs behind why we're able to be able to accomplish this, but also have customers talk about it as well. And in addition, as a part of being a Gold sponsorship, we have a breakout session with DSW, designer shoe warehouse, great, you know, get stores here all over Massachusetts. Or 400 locations throughout North America. Getting on stage and talking to us, talking to the audience about their experience with Infinidad and the Infinibox and how happy they are with that solution. And that's the South Pacific Jay. If you show up at the show, please register. I think it's going to be a great talk. It's Wednesday at 11 a.m. Wednesday at 11 a.m. Right, yeah, and we're also, sorry, we're going to be doing a lot of fun things around the booth. We've got some interesting giveaways. We're going to be doing some social promotion. It's going to be fun. We're going to make a lot of noise. We have a packed AR schedule. We're doing a lot of briefings. We're, it's going to be fun. We're going to be joining you guys on theCUBE. That's right, and theCUBE will be there. We got a double set again, and so that's going to be awesome. And I think, you know, from my standpoint, just personally speaking, having these conversations really opens up, you know, your eyes and your mind. People are making big bets right now. And if you're going to make a big bet, whether it's betting on all flash or betting on some kind of hybrid or betting on a new architecture, talk to the people who are doing it and learn. So I say, thank you guys for coming in today. You've really helped us with our thinking, with some of our forecasts, challenged us to go talk to some of the drive manufacturers, which we've done in some interesting sessions, as you can imagine. And as I say, Florie's new forecaster coming out this month. And really appreciate the time and the collaboration. So Randy, Steve, great to see you guys. Thanks for having us. All right, thanks for watching this special presentation. This is Cube Conversations, this is Dave Vellante. We'll see you next time.