 From the SiliconANGLE Media office in Boston, Massachusetts, it's theCUBE. Now, here's your host, Dave Vellante. Hi everybody, this is Dave Vellante. Welcome to this CUBE conversation with Dr. Rico, who's the CMO of Infinidad. It's still, I still have a hard time saying that, Dr. you're an engineer, and I love having you on because we can talk storage, we can go deep, and we can talk trends and marketing trends too. But so welcome, thanks for coming on. So tell me what's new since the scale to win launch that you guys had, tell me what's, is everything shipping now, what's the uptake been like with customers and the reaction? Yeah, the reaction has been phenomenal. This, as you may recall, you were there, it was the biggest launch in our history, which was fantastic, and the reaction has just been overwhelmingly positive with customers, with partners, with analysts, even in some cases with competitors, has been interesting. We had a lot of things that were already shipping, they were an early customer release, there were a few things that we had started shipping in December, and the things that we said would be coming in 3Q, we GA on time, so they're now all generally available, except the stuff that we talked about, that would be available in 2020, which right now looks like it's on track, it's doing very, very well. So VMware, VMworld is coming up later on this month. Things are obviously changing, there was an announcement recently that VMware is going to acquire Pivotal, so a little bit of financial engineering going on, stock rose 77% on the day when the Dow dropped 800, so okay, funny money, but things are changing in the VMware ecosystem. You certainly saw, this is our 10th year at VMworld, we go back and you hear Todd Nielsen back in the day talk about for every dollar spent on a VMware license, 15 was spent on the ecosystem, we're kind of delizing VMware now, which is sort of interesting, but I'm curious as to what you're seeing, what that all means to you, I mean still half a million, 600,000 customers, you've got to be there, you guys have great success at that show, so your thoughts, what's going on at VMworld this year? Yeah, kind of loaded there, and first of all congratulations on the milestone, that's great, 10 years is super, I remember probably seeing you at the first one there, of course we knew each other longer, and sure I get the incestuous money changing of hand there, I think it's good in one respect, you certainly see VMware making big inroads with VMware on AWS, and this now with Pivotal will be a good launching platform for Dell, as well as VMware to be a little bit more in control of their own destiny, and it's certainly the way a lot of people are going, we're doing a lot of that ourselves, not so much in a sense we don't have a cloud platform that we sell as a total encompassing platform, but of course with Nutrix Cloud and the big players, and then certainly a large portion of our customer base, our cloud service providers, they love our stuff, it helps them compete, it actually gives them in some respects a competitive advantage, but VMworld itself, lots going on there, we have amplified our presence once again, because VMware does represent a large portion of our customer base, so we're very proud of that, we're very proud to be a technology-aligned partner of VMware's and we're expecting to see a really good show and a really good cloud crowd as they return back to their home base in San Francisco. For us this year, it's gonna be a different experience, we're telling more of the software story, more of the portfolio story, more about how you scale to win, we have a virtual presence this year which is going to be very helpful in telling that story, customers can come in and they can see more than just a box that in our world is really not important, because for us it's all about the software and stuff we do, we have an in-booth theater, we do have some private meeting spaces as well to take people into a bigger, deeper drill down, but the virtual experience will allow them to touch and feel stuff that maybe they didn't get to do before, and that's gonna be kind of exciting as well. So you mentioned CSPs, we had Michael Gray, a thrive on a while back, and he was saying that, look, he likes your product because it allows him to do other things, you don't have to worry about the old tuning and managing and able to reshift labor. I felt like that was an interesting discussion primarily because you've got all these cloud service providers that everybody thought AWS was just gonna kill, and if anything, it's elevated them. What are you seeing in the CSP space? Yeah, Michael had a lot of interesting things to say, they definitely love the fact that we enable multiple workloads without them having to do lots of cautious planning and replanting and shifting and shuffling, and we are seeing CSPs becoming more value-add to a lot of businesses, especially the mid-market and the smaller enterprise, where people may want more than just infrastructure, they need that application level support, and companies like Thrive and some of our other really good customers, US Signal, and they're all capable of, Flexential's another one, they're all capable of providing services beyond the hardware, they're capable of providing that application support, the guidance, and then the case of Thrive, the cybersecurity guidance especially, which is really, really critical. So they're growing, and they're also, by the way, working with AWS and Google and Azure to provide that capability as well when necessary. Well, that leads me to the sort of multi-cloud discussion. In our industry, we tend to have this alphabet soup of acronyms, another reason I like talking to you because we can kind of cut through that. And I love the marketing, I think marketing helps people understand what's going on, differentiate, it gives you an indication of where the industry is going, and multi-cloud is one of those things that, I mean, I've kind of said it's a symptom of multi-vendor, and it's more so than a strategy, but increasingly it seems like it's becoming a strategy with customers, and you just gave an example of Thrive working with multiple cloud vendors, clearly VMware wants to be in that business. What are your thoughts on multi-cloud and hybrid? What does it mean for Infinidat, and what's your strategy there? You know, it's interesting, because I just read an article the other day about the definition of multi-cloud and whether it's being abused, and you know, I look at it as someone just trying to tell their story and give it some favor. I think at the end of the day, every business is going to be talking to multiple platforms, whether they want to or not. You know, there are many customers and companies out there, businesses who are in our customers, who have gone the way of the cloud and repatriated certain things, as they've found that it may work, it may not work, and there are many cloud providers who are trying to do things to accelerate migration of applications, because they see that certain applications don't work. You know, we've got one of the cloud providers buying a NAS provider, another one buying very recently an NVMe-based flash company to try to pick up those loose workloads where they might struggle today. But at the end of the day, everybody's going to be multiple and whether it's because they're using cloud services from a software perspective, or whether they just need to basically broker and maintain sort of that independence so that they can maintain some cost control, availability control, security control, and in some cases, it will remain on-premises and some of the things will be off, just so that they can get the applications closer to their end users. So, you know, what is multi-cloud? Multi-cloud really is just one of those terms that literally means what it says. It's your business running in multiple places. It doesn't have to necessarily be simultaneously by the same application. A big part of your value proposition is the simplicity. We've heard that from your customers and you guys obviously pushed that out there. I want to ask you, because you mentioned repatriation and you know, cloud keeps growing like crazy. And the on-prem, you know, not so much. You guys, a smaller company, you're growing, you're stealing share, so maybe it's that simplicity thing. Here's my question, so it's around automation. The cloud providers generally in Amazon specifically have driven automation, they've attacked the IT labor problem and they're able to charge for that. And so my question is, are you seeing that you're able to attack that IT labor problem in a similar sense and bring forth the value proposition to customers is look, we can create a cloud-like experience on-prem. If you want to go to cloud, great, but if you want to stay on-prem, you're going to get the benefit of being able to shift resources to more strategic things and not have to worry about all this heavy lifting. Are you seeing tangible evidence of that? We're seeing significant tangible evidence of that. And you know, a couple of things, you know, you talk about growth, right? I think when we did the launch, you know, only a few months ago, we were at about 4.6 exabytes of capacity shipped and we just passed 5.1. I mean, that's some significant growth in just a few months. It's like a 33% growth just from the same time last year which is fairly significant. And of course, if you're familiar with the way we talk, you know, you have an engineer as the head of marketing, we like to tell the truth. We don't like to mask too many things and confuse people. We don't like talking about effective storage because effective capacity doesn't really mean much to some people. So that's, you know, this is what we shipped and it's growing rapidly. And a lot of that is growing in part because of the significance of the message and in part because of this need to control costs, contain costs and really operate in a more modern way. So get back to your comments about cloud and cloud operation. That's really what people want. People like the consumption model of cloud. They don't always like the cost and the hidden costs. So simplifying that but giving them the flexibility to have either an OPEX or a CAPEX that allows them to grow and shrink as they move workloads around because everybody grows. Even on-prem is growing. It's just, you know, it's the law of numbers, right? Cloud is growing, absolutely, but on-prem really is growing. And then the other thing they want is they want the operational flexibility and that's what we talked about in our elastic data fabric. They don't like constantly having to rejigger and rebalance workloads. InfiniBox by itself, the platform of InfiniBox takes away a lot of that mystery and magic because it kind of hides all of the complexity of that workload and we take the randomness out of the IO. I think maybe Craig Hibbert mentioned in his video as he was describing in detail how that happens. And I remember Michael Gray talking about that as well. So those things come out in a single InfiniBox. But even if you said, well, I still want to move my workload from this data center to an adjacent data center or perhaps a data center in another facility, excuse me, another city, so it's closer to the end user, making that transparent to the applications is critically important. Yeah, so you talked about growth. I mean, it's about a half a petabyte, I mean, sorry, half an exabyte in just a few months, a couple months really, that's growth. But I want to ask you about petabyte, petabyte scale is kind of key to your marketing. There are companies that don't do that in a year, Dave. Exactly, so petabyte scale is big party of marketing. Two questions, why is that relevant or is that relevant to VMware customers and why so? And then does it scare some people? Oh, you asked a great question. It absolutely scared some people. And I know that there are some pundits out there, industry pundits who basically don't agree with our messaging. But this is the business problem that we targeted to solve, right? There are a lot of people out there who don't think they're at petabyte scale yet because maybe their individual applications aren't petabyte scale, but when you add it up, they get there. And a lot of our customers, our existing customers didn't start with Infinidat at petabyte scale. They started a couple hundred terabytes perhaps. But they're at petabyte scale now, in fact, over 80% of the customers and systems that we have out there today are above the petabyte. We have customers that are in the tens of petabytes. We have customers that are in the hundreds of petabytes. They grow, they grow rapidly. And why is that? Well, two factors, really. Number one, if you go back to probably when I first met you, back when I had your hair, at least in quantity, we were kind of cresting that terabyte mark, right? And what was the problem? The problem was nobody could figure out how to deal with the performance. Nobody wanted to put that much risk on a single platform, so they couldn't deal with the availability. And they really didn't know how to deal with even the serviceability at that scale. So terabyte was a problem solved 25 years ago. And then things grew rapidly from there. Now we're at the same juncture, just three orders of magnitude later, right? Well, that's interesting because you're right. People didn't want to put all that capacity under an actuator that caused performance problems. They were concerned about just availability. And then two things happen, sort of simultaneously. Flash comes along and you would say it was sort of a bandaid to some of the performance problems. And then you guys came up with like this magic sauce to actually use spinning disk and get the same performance or better performance you would argue with than flash. And so as a result, you were now able to do a lot more with the data, the concerns about that much data under the actuator somewhat attenuated because I mean, you've got now so much data. You've got to do something with this flywheel effective. You've got tons of data, machine intelligence and AI now coming into the picture. You've got cloud, which has been this huge tailwind for the industry and for data creation in general. And so you see like the IDC numbers and forecasting growth of data and storage, it could be low. I mean, the curve could be bending kind of more than exponentially. Your thoughts on that? Yeah, it's an interesting observation. I think what it really comes down to is our storyline is math is greater than media, right? And when you look at the flash being the panacea to performance, it was just a step in the evolution, right? You can go back and say spinning disk was the same solution to the performance problem. 20 years ago, 25 years ago even, it was 5,400 RPM disks. And then very rapidly servers got faster, the interconnects got a little bit faster, they were still mostly differential scuzzy. Then it was 7,200 RPM disks. And I promise you by the way, that if you're running 5,400 RPM disks, you install 7,200 RPM, all your performance problems will go away until the day you install it. And then it was 10,000 RPM disks. And then it was 15,000 RPM disks. And it still wasn't getting fast enough because you went to fiber channel, one gig fiber channel and then two gig fiber channel, four gig fiber channel, eight gig fiber channel. The interconnects got faster, the servers got faster, there was more cash on the servers. Then this thing came along called solid state disk, right? And then it was SLC, single layer cell technology. But don't worry about it, it's very expensive, not a problem, you only need 4% of your application. Oh, no, no, I'm sorry, 10%. No, I'm sorry, 30%. Oh, what the heck, MLC is now a little bit more reliable, so let's just make it all flash, right? So that was the end of the story, right? No, servers continue to get faster, the media continue to get faster and denser, right? So now the interconnect isn't fast enough, so NVMe is the answer to life, the universe and everything. All right, well, wait, I got a better answer for you. It's SCM, the storage class memory. In parallel with that, by the way, there were some vendors out there who said that's still not fast enough. We wanna put a bit more DRAM in the servers and do things in memory, we went in memory databases. I guarantee you, whatever you do from a media perspective, and this is my personal guarantee to you, it's obsolete by the time you're up and running. By the time you get your applications migrated, configured, and running with business value, it's already obsolete, some vendors got something better coming out. The right answer is the stuff you talked about, the right answer is everything that you're doing for your business apps. It's AI, it's ML, it's solving the problems in software. And you said we use disk and make it fast, it's not disk by itself, right? It's DRAM, it's a lot of DRAM, which by the way, is orders of magnitude faster than flash, than NAND flash. And even if it's SCM, it's still orders of magnitude faster than that. What we use the disk for today in the architecture is the cost factor. We take the randomization out in the flash and in the DRAM, and we use the SAS and the back end to manage cost, but we use it in a way that it performs well, which is highly sequential, massively parallel. And we take full advantage of that back end band with to do that, with that massive DRAM front end, our cash hit ratios are unparalleled in the industry, and we use it even more effectively that way. But if our architecture already evolves, so if SCM becomes more stable and becomes more cost effective, we can replace that SSD layer with SCM. And if the economics of QLC or something beyond that come down, we'll replace the back end with that. Do you ever look at what you're doing today as sort of a modern day symmetric? I mean, a lot of things you just said, I mean, you've got a lot of memory. You've got a massive back end. Those were two of the characteristics of symmetrics. Now, of course, fast forward, whatever, 30 years. But a lot of it was sort of intelligence and understanding sort of how data works. Is it a fair sort of evolution, or is it radically different in terms of mindset? I mean, I know the implementation is right. Yeah, I mean, it's not an unfair comparison. I mean, tiered storage was around before symmetrics, right? So it certainly existed then too. It was just, at the time, it was a significant innovation. Of course, two layer at the time, right? So a big cash front-ending, some slower media, and then taking advantage of the media on the back end. The big difference today is that if you look at what symmetrics became through its evolution, DMX and VMAX and now PowerMAX, it's still tiered storage. You still have some cash that's front-ending, some faster media with PowerMAX. You're dealing now with an SSA back end. But what happened with those types of architectures is the tiering became more automated, but you're still moving information around. You're still moving information from one set of disks to another set of disks later in the cycle. You're still trying to promote things to the cash upfront. We're doing it in real time. We're doing it by analyzing the data on the way it comes in. We're reassembling it, again, taking the randomization out. We're reassembling it and storing it across multiple disks in a way that it increases our probability of pulling that information and associated information back when we need it later. So there's no movement. Once it's placed, we don't have to replace it. It's already associated with other data that makes sense, and that gives us a lot of value. And the secret sources, or the outcome of the secret sources, you're able to do that very efficiently. Historically, you haven't been able to yet do a lot of garbage collection, a lot of data movement, and that just kills performance. There's really no garbage collection necessary in our world. We also use very modern data structures, our patents. A lot of them on our neural cache deal with the fact that we use a tri data structure. So we're not using old fashioned hash tables and LRU algorithms. It's very, very rapid traversal of these trees. And you're taking advantage of machine intelligence inside the software architecture. That really is sort of the new innovation that really wasn't around to be able to take advantage of that 20 years ago. Maybe it was, it was just not cost effective really to do. The math was there, we'll put it that way. The math was there. I mean, there's been lots of evolutions of that over the years as well. But we continue to evolve and innovate. And one of the cool things I think about working in Infinidad is the multiple generations of engineer where you've got people who understand that math. They understand the real nuances of what it means to operate in the world of storage, which is quite a bit different than operating say in networks or CBUs because data integrity is paramount. There's lots of things that go on there as well. But we also have younger generations, generations who like new challenges and like to reinvent things. So they find newer and greater ways to do things which is exciting. Some systems thinkers, and I don't mean server thinkers. I mean people who understand systems design and all the way through and newbies who are super smart and like to say want to learn. And solve problems, go back to the petabyte scale discussion, solve problems at petabyte scale, right? Even if the customer doesn't need that necessarily, to solve that problem is critically important. Because even if you look at, let's just take NFS for example, right? Most NFS systems deal with thousands of objects, hundreds to thousands of objects. Our NFS implementation deals with billions, right? Do you need billions? How many applications do you know that have billions of objects? But being able to do that in a way where performance doesn't degrade over time and also do it in a way where say our NLM implementation isn't impacted by any type of service events. We can take a node out and it doesn't impact NLM. There's no degradation in performance. There's no impact or outage in service, right? All that's important even when you're dealing with smaller application sizes because they add up. They really do add up. You also brought up the point about density and actuator density, right? Back 25 years ago when we were dealing with the first terabyte storage system, how much storage did you have on your laptop? How much you have today? Right. You're probably more than a terabyte. They were laughing about putting terabyte on the floor and now you got more than a terabyte on your laptop. Things are changing. Yeah. I want to ask you where you see the competition. We talked about all flash. We've had a long conversation and many conversations in the past about this. But you really, you know, the all flash, you kind of described it as a band aid, essentially my words, but it was sort of a step function. Okay, great. You have one company really who achieved escape velocity in that business in terms of pure. But is that where you're seeing competition? Are you seeing it from, you know, the hyperscalers? Where are you? Yeah, you know, it's interesting. You know, you look at companies like, you know, we admire what they do, especially with regard to marketing, they do a really good job of that. They also have some really interesting ideas in innovating the media, which is great. It helps us in the long run as well. We just look at it as a component of our system, not the system, which makes it different. We don't really see the AFA, you know, the small scale AFA as the majority of our competition. We do run into them, but typically at the lower end of the opportunity. Even within the bigger companies, you know, that have competitors to those products, we run into them in smaller opportunities, not bigger opportunities, or we run into them where there's a significant performance advantage as long as you don't mind the scale out approach to solving the problem. Unfortunately, when you're using AFA's to scale out, you know, you're putting all of the intelligence requirements on some poor storage administrator or system administrator to figure out what goes where, right? We take all of that away. So once it starts to scale, that's where we come into play. And we don't see tons of competition there. Certainly we're seeing competition from the clouds and the competition from the clouds is more born of customer mandates and company mandates. Sometimes I'm not quite sure that everybody knows why they're moving to the cloud and what problem they're trying to solve. But once they start to see a story that says, hey, if the reasons are, and you do understand those reasons, if the reasons are agility and financial flexibility and operational agility, as well as acquisition agility, you know, we have answers to that. And it starts to become a little bit more interesting and compelling. All right, one of the highlights of VMworld each year is your dinner, your customer dinner. I crashed it a couple of years ago when there were no other analysts there. And then last year, again, it was in Vegas, so you chose a nice steakhouse. This year we're in San Francisco. But I had some great conversations with customers. I remember speaking to one customer about juxtaposing VSAN to Infinidat's platform. And the difference, VSAN's taken off, doing really well, but he helped me understand kind of the thinking from their standpoint of how they're applying it to solve problems and why VSAN wasn't a good fit your system was. That was just one of many conversations. Last year had some, again, other great conversations with customers. What are you doing this year? You having a customer dinner? What's that, can I go? We are? Yeah, we're, love to have you again, Dave. The invitation is there. I got shot there. Yeah, the invitation is definitely there. You know, a couple of years ago, we didn't invite analysts. And you know what, it was a mistake. We learned that lesson, and to a large part, we credit you for showing us how wrong we are. Our customers are very loyal. They're some of the most loyal in the industry. I mean, don't take my word for it. Go on a Gartner peer insights and look at our numbers compared to everybody else's. Any pick a vendor, we're at the top of the list with regard to not only the ratings, but more importantly, the customer's willingness to recommend in every category too, by the way. It's not just product quality and performance and it's service support, it's ease of doing business. It's an entirely different experience. So we love having the customers there. And the customers love having you there too. They love having you and your peers in the industry there because they love learning from you and they love answering the questions and getting new insights. And we'd love to have you there. We're going to be at the Mint this year, San Francisco Mint, not the current one that's printing coins, but the original historical site. And we have invitations out to about 130 people because there's only so much room we have at the event, but we're looking forward to a great time and a great meal and good conversation. That's great. Well, VMworld is obviously one of the marquee events in our industry. It's the fat middle of where the IT Pro goes and we're excited. It used to be Labor Day started the fall season. Now it's VMworld. Well, Doc, we'll see you out there. Thanks very much for coming on theCUBE. Good to see you again. All right, excellent. All right, thank you for watching everybody. This is Dave Vellante in theCUBE. We'll see you next time. And we'll see you at VMworld 2019.