 From Austin, Texas, it's theCUBE, covering Pure Storage Accelerate 2019, brought to you by Pure Storage. Welcome back to theCUBE. Lisa Martin, Dave Vellante is my co-host. I'm at Pure Accelerate 2019 in Austin, Texas, and Dave and I are really pleased to welcome to theCUBE for the first time. John Calgrove, CTO and founder of Pure Storage, calls, welcome to theCUBE. I'm glad to be here. Thanks for having me. And happy 10th anniversary. So 10 years ago, I'm sure you couldn't have envisioned 3,000 people Austin being taken over by a sea of orange. But let's go back 10 years. Why did you found Pure Storage? What did I found it? Well, I wasn't really ready to be retired yet. Flash, you know, I sort of have seen from, when I worked at Amdol many years ago, all the way through Veritas, I saw disks continuing to get bigger and bigger and effectively slower and slower. Because when they don't get any faster and they get bigger, they get slower for everybody of data. And Flash was a catalyst that was going to change that. But it was the catalyst. What we really wanted to do was to completely change the storage industry. Everything that had annoyed me about the storage industry through all the years in Veritas, all the complexity, you know, all the bad customer practices that the industry forced on people, I wanted to change all that. You know, think of what you demand from your personal tech, from your iPhone or your laptop or your tablet. Customers should demand that kind of quality, serviceability, ease of use from their enterprise ITger. You know, when I started my career in the early 80s, I was at IDC and they didn't have a storage analyst. And I started following mainframes. And I learned a lot about channel command words and IO subsystems and it came to the conclusion that this is a really hard thing, hard problem to solve. And so I got interested in it. You obviously did as well. I'm interested in, when you went from Amdol to Veritas, you had to do some unnatural acts with software to make IO better because of the spinning disk and understanding the latencies and the chatty protocols and everything else. When you went and thought about Pure and when you think about great architects and I've obviously put you in that category, you chose Flash. Others, like another great architect, Moshe, have said, you know what? I can even squeeze more out of spinning disk. What led you to Flash versus trying to squeeze more blood from the spinning disk stone, if I can phrase it that way? I think I tend to be more of an extremist on things like that. And I think that's been the key to Pure's success. We were not the first all-Flash startup. We were the first to focus on affordable Flash. If you're going to change the world, you have to make something for everyone, not for an elite few. But the other thing was we were all Flash. There were a lot of other startups that were hybrids that were squeezing more out of the disk and we just went all Flash from the beginning. Everything about us is all Flash. So as the future goes more and more towards Flash, we're in a stronger and stronger position. And you think that was the game changer that led Pure to be that unicorn that I peed four years ago versus those other startups who were trying to do similar things with Flash? So that focus helped us a lot with that. You know, the biggest thing that, as I said before, Flash was a catalyst. The biggest thing we've brought to the industry is the simplicity and the evergreen business model. And it's really cool to see all the big companies that we've competed against all these years mimicking a lot of that. But that's the differentiator. Flash was the catalyst that lets you do that. Well, so I'm interested as a little bit of an industry historian in some of the factors that led to your ability to achieve escape velocity, which used to be defined as an IPO. I mean, I would argue the three-part achieve escape velocity. It was a $250 million company before it got acquired for $2.5 billion or whatever it was. Never reached the billion, never even came close. You were the first storage company since NetApp to achieve, you know, $1 billion revenue. And you're well on your way to $2 billion. You do probably 1.7 this year. In addition to what you've said, are there other factors that we should consider in our B-School case study on Pure? You know, I think one of the things we've tried to do is we've tried to build a company that's going to be in it for the long-term. So we never wanted to settle for an acquisition. We want to build a long-term, enduring, great brand. And part of that, you have to build more of a partnership with your customers. You have to be a good partner to your partners. Right, if you are short-term focused, if you try to squeeze every dollar you can out of people, they don't like you, they don't want to come back. If you build something great and you partner well with the environment around you, you can build something long-lasting. And we wanted to do that from the beginning. We focus a lot on culture and things like that to help us do that. Well, it's impressive. Congratulations, everyone, because 3PAR couldn't do it, compelling couldn't do it, Isilon couldn't do it on and on and on. And EMC at the time was really, it's about EMC, that's how you went after. They were able to do virtualization and freeze the market on 3PAR. They were able to do a low-cost, call it the compelling killer. They were never able to figure out, now maybe they got distracted with Elliott management and everything else, but they were never able to figure out how to squash you guys. And that's impressive that you're able to live through that. I mean, one of the things we've always tried to do is be supremely disruptive and that does make it harder for them. So I got to challenge you on a couple of things that have come out largely from your competitors, but I want to get your take on it. The first one is scale out. Well, how come pure doesn't scale out? I'll leave it there. I have my own thoughts that I've shared with Lisa, but two controller design. One thing I'd point out is, well, FlashBlade, one of our products is scale out. FlashArray, our first product is not scale out. Scale out isn't a capability for a customer. It's an architecture in how you build the product. When I scale out, I have more complicated software. I have more components. More components lead to more failures. If I have a piece of memory and it's going to fail at a certain annual failure rate and I have 10 pieces of memory, I'm going to fail at 10 times that same rate. So scale out introduces complexity, it introduces more components, and then you have to say, what do you get from it? So if our customers needed a lot more performance than we're delivering, if they needed a lot more scale than we're delivering in the FlashArray product, we'd then react to that and go build scale out. Where the FlashArray sells, we don't see that as a major market need, it's more of a niche. Where FlashBlade sells, then there is much more of a need for that, and that's why FlashBlade was scale out from day one. Well, am I correct too? The other thing you get from scale out is non-disruptive controller swaps, but you've solved that in other ways, right? You know, you say you get non-disruptive controller swaps, I will point out that if you look at these scale out architectures out there, there's a set of them that do provide that, but actually the larger set of them don't provide it, because what they're doing is they're making what they view and what the customer views is one monolithic array built from a set of scale out components. So in those architectures, you can't swap out one part of the scale out, you have to swap out the whole thing. The other thing I heard that I love this analogy is you don't really see planes anymore, you see them, but you really don't want to fly them because they're old with four engines versus two engines because the two engine planes are so much more reliable. All right, the other question is on proprietary Flash modules. You guys have chosen your philosophies, do things that you can't do with just off-the-shelf components, so you've gone proprietary, and there's history there. I mean, three par with custom ASICs, but I'd like you to share with us your philosophy on what you're doing. So there's a couple of dimensions to that. Number one, we have gone with proprietary Flash modules, but in our Flash array, we could plug in off-the-shelf drives anytime we want. And in fact, today, our XR2 line, the lower-end models use off-the-shelf Flash, and the higher-end models use the proprietary. What we get with the proprietary is our own firmware on there. It's the same NAND chips, the same NAND controllers. It's all the same components, but it's our firmware. And our firmware only has to support one application, our purity operating system. However, the customer reads and writes data into the array, we write it the same way down to the Flash, we read it back the same way from the Flash. So by making simpler firmware that only has to solve that one problem, we get better performance out of the Flash, we get longer life out of the Flash, and we get order less than one third of the failures of Flash drives. Now, the Flash drives we were using were already failing a lot less than disk drives, but we've gotten better than three times the reliability by going to our own Flash modules. Tearing, your philosophy on tearing, five, 10 years ago, there was a big thing on automated tearing, we're going to put the hot data on the high performance, either disk or flash, and the slow data on the cheap stuff. Your philosophy on tearing, you know, I think, I infer, you don't believe in tearing. Why not? Or maybe, I don't know if I'm worth your time. So tearing is another thing that adds complexity. So, why do you tear? You tear because you say, oh, I can't afford all of the better thing, so I'm going to layer it in with something that's a little cheaper. If you can get by without tearing, that's a better solution, it's a simpler solution. Simplicity is a theme here. The CompuVarid Acquisition, your file system guru, to my knowledge, but I've read about it, a strong file system. What do you intend to do with that? It's concerned about it forking your existing products. How do you respond? So, the CompuVarid file system, we're going to put that on top of our Flash Array line and make that a unified architecture where you can support block and file. CompuVarid is a very complete file protocol stack and, you know, file protocols are a lot more complex than block protocols. Implementing all of the SMB protocol is not an easy thing. It takes a bunch of time, so it's a way to accelerate that and get a very complete protocol stack for that product. FlashBlade will continue on with its own scale out file protocols, file and object protocols independent of that. Last question I had is on, there's some criticism that's been laid on you guys on the Evergreen, the controller performance of controller upgrades, which I've not heard, we didn't hear that from customers, we've asked some customers that, but I'd love to get your take on why is there no guarantee of performance improvements as you go to subsequent controller swap outs, your thoughts? So, what we guarantee is you'll get the like or better. So, you might get a new set of controllers that are performed about the same, you might get one a little better. Generally speaking, every time we've done it so far, it's moved to better. It doesn't move to radically better, but it moves to the better. So, we are guaranteeing that. It's just a question of how much do you choose to deliver with that? What you're doing is you're keeping the array new. It's not so much about making huge strides in the performance, it's about keeping the array new. But there's another nuance there that I want to test. It just conceptually seems to me because the way you ship software constantly, that you're making incremental improvements throughout that three-year period. First of all, is that an accurate assertion? It's actually very accurate. The first time we started really looking at how much better, we realized that we had moved the needle on the old gear about, I think it was about 60% up during the time period. So, yeah, there was sort of a little less gain. So, the proper measurement is, okay, what's the performance from day one delta to the controller upgrade? That is more significant versus the controller swap, day, whatever, N to N plus one, if that makes sense. I think both are valid ways to look at it. The biggest thing is the customer doesn't have to migrate and the migrations are the most horrible event in storage, right? It's like moving your house, for everyone who's moved, you got to pack everything up, things get broken, things get lost, it's just a mess. You don't have to do that and the array just gets bigger, denser, more power efficient. It gets better and better over time and you're on that forever. We are happy to do controller swaps after three years, six years, nine years, 12 years. We will continue to do that as long as customers are paying for that, it's our job to keep improving it and to keep making it better. We've done a lot of research on array migrations at a minimum, your auntie to do an array migration is $50,000, that's what our data shows. We talked to a very large practitioner last night, he said, when I'm doing an array migration, I start six to eight months ahead of time because it takes that long to do an array migration. Array migrations are horrendous and anything you do to avoid those, it's worth it. That's all I had, that's awesome. Thank you for addressing those questions. The acceleration, pun intended, that Pure has achieved in its first 10 years, we talked about customers all the time. We've had a number on yesterday from law firms to utilities to F1, we'll have more on today, but in order to achieve what Pure has, you have had to build a culture that's pretty unique. One, this vibrant orange color that just screams energy, boldness. Two, we're in Austin, Texas, Dell Technologies Backyard. Give us a little bit as we wrap here about how you and your co-founders have developed and really fostered this culture of passion that is delivering more than your competitors would like to see. Well, so one of the things that was a key part of the culture is we didn't just hire a bunch of storage people. We had a few early on because you need some experience in the history, but an awful lot of the people we hired came from other backgrounds, other engineers, marketing people, et cetera that did not come from storage. And what we challenged people to do when they come in the door is, we're hiring them because of their brain power, right? We don't own mineral rights somewhere, we don't have buildings, we don't have a lot of assets. Our asset is our people and what they can produce. And obviously, if you think back, well, when I was the only employee, right, I was doing every job. Ideally, everyone we've hired since can do whatever we've hired them to do a lot better than I could do it. And that's a philosophy you want to keep going. Every person at Pure should be focused on using their brains, using their creativity to deliver the most value possible to disrupt things where they can to always look for how we do things better and to always be looking to hire better than them. So it kind of gets into the next 10 years. Don't hate me for saying this, but in retrospect, the first 10 years, you had it kind of easy, you caught EMC off guard, you drove a truck through their install base, NetApp, Miss Flash, you guys executed obviously, we talked about that billion dollar company. The next 10 years, a little different. Where's the TAM expansion come from? In the next 10 years, it's multi-cloud, it's new AI workloads, it's lower cost solutions that get you more of the market. It's partnering with backup, but you got cloud, you got competitors that are starting to figure it out. How do you see the next 10 years to go from beyond where you are and that next hike? Well, so I'll start by saying, when you start a company, you dream of success. And the first 10 years have been as good as you could possibly have. No doubt. So, A, hopefully the next 10 years will continue that way. I think you touched upon one thing is the cloud. People have been through the hype cycle of saying the entire world's going to be cloud, there's only going to be three data centers in the world and it's going to be Amazon, Microsoft and Google. They now understand the cloud is a tool and you need to use it properly. So, one of the focuses we're going to be working on over the next several years is making sure that someone can have their data, their application on-prem. They can decide I want to put it in the cloud, move there seamlessly, move there as easily as you move from one of your cell phones to the next model. Move from one cloud to another cloud, move from that cloud back on-prem. Whether you want to move the data, the applications, both, and get the same kind of service, the same kind of experience. So that's going to be a big thing. Yeah, a lot of work to do there, but yeah, but there's an opportunity, isn't there? It's the way everybody wants to run, it's the way everybody should run. Running an IT service to deliver value to your company, value to your organization should not be rocket science. And our job at Pure is to make that accessible to everybody so everybody can deliver that kind of quality experience to their organization. Ken, it's an obvious question, but you see that as technically feasible over the next five to 10 years. Yeah, it is technically feasible. This goes back to one of the things that I was mentioning before, with Flash as a catalyst, one of the things Flash helps do to make this simpler is it frees you from the geometry constraints of disk. You don't have to care as much. Another thing that's making it possible is faster networking, right, and better networking. Then again, you have all the compute and GPUs and co-processors and things, pushing things. As you get to where resources are more plentiful, then you have the ability to trade off some of the, I've got to get every microsecond out of this thing for the simplicity for that ease of use. And that lets you deliver something better in the long run. If I perfectly tune something, I might be able to do a little bit better, but I'm not going to be able to keep it in tune and I'm going to spend my whole life re-tuning it and re-tuning it and finding it out of sync. Simplicity that drives so much efficiency, agility that drives so much value. Well, Kauze, thank you so much for joining Dave and me on theCUBE this morning from Accelerate Day 2. You talked about Flash being a catalyst. That sounds to me like Kauze has been one of the major catalysts of Pure's success. Happy 10th anniversary. We look forward to the next 10th. Thanks a lot and thanks for having me. For Kauze and Dave Vellante, I'm Lisa Martin. You're watching theCUBE from Pure Accelerate 2019.