 Hello everyone, welcome to this special CUBE Conversation. We're here in Mountain View, California, at the Pure Storage headquarters. I'm here on Castro Street, one of the many buildings they have here as they continue to grow as a public company. Our next guest is Kix, Vice President of Strategy, employee number six at Pure. Great to see you, thanks for spending time. Thanks for having me. So cloud is the big wave that's coming around the future. It's up to here now. People are really impacted by it. Operationally, coming to the reality that they got to actually use the cloud as benefits for it, or many multiple benefits. But you guys have major bones in storage, flash arrays, continuing to take territory. So as you guys do that, what's the cloud play? How do customers who are using Pure, and we've heard some good testimonials. You got a lot of happy customers who've seen great performance, easy to get in, reliability, performance is there on the storage side, on premise. Okay, now operation says, hey, how to build faster. Cloud is certainly a path there, certainly a good one. Your thoughts on the strategy for the cloud. Absolutely. So look, we're about 10 years into the journey here at Pure. And a lot of what we did in the first 10 years was help bring flash onto the scene. And we had a vision when we started the company of the All Flash Data Center. And I like to first of all remind people that look, we ain't there yet. If you look at the analyst numbers, about a third of the storage sold this year will be flash, two thirds disk. So we still have a long way to go in the All Flash Data Center and a lot of work to do there. But of course, increasingly customers are wanting to move workloads to the cloud. And I think the last couple of years I've almost seen a pendulum swing a little bit more back to reality. You know, when I met with CIOs two, three years ago, you often heard we're going all cloud or we're going to be cloud first. And you know, now they're a few years into it. And they've realized that cloud is a very powerful weapon in their arsenal for agility, for flexibility, but it's not necessarily cheaper. And so I think the swing back to really believe it in hybrid is the model of the day. And I think the other thing people have realized in that journey is that the cloud really works best when you build an app for the cloud natively. But what if you have a bunch of on-prem apps that are in traditional architecture? How do I get them to the cloud? And so one of the things we've really focused on is how we can help customers take their mission critical applications and move them seamlessly to the cloud without rearchitecture. Because for most customers, that's where they're going to start. They can build some new stuff in the cloud, but the bulk of their business, if they want to move substantial portions to the cloud, they've got to figure out how to move what they've got. And we think we can really add value in that. And the economics of the cloud is undeniable. People who are born in the cloud will testify of that. Certainly, as you guys have been successful on-premise with the cloud, how do you make those economics be seamless as well as the operations? This seems to be the number one goal. Can you talk about how important that is and how hard it is? Because it sounds easy to just say it. But it's actually really difficult to have seamless operations on-prem because, you know, Amazon, Google, Microsoft, they all got compute and storage in the cloud. And you got storage on-premise. This equation is a really important one to figure out. What's the importance and how hard is it and what are some of the things that you guys are doing to solve that? Yeah, so I heard two things in that question. One around cost and one around operations. And, you know, the first thing I think that has been nice to see over the last couple of years is people realizing that both the cloud and on-prem are cost-effective in different ways. And I think a little bit of the way that I think about owning a car. Owning a car is relatively cost-effective for me. And there's times when taking an Uber is relatively cost-effective. And I think they're both cheap. When you look at it on one metric though about what I pay per mile, it's way more expensive to take an Uber. If I look about the acquisition cost, it's way more expensive to own a car. And so I think both of them provide value in my lives in the way that hybrid does today. But once you start to use both, then the operational part of your question comes in. How do I think about these two different worlds? And I think we believe that storage is actually one of the areas where these two worlds are totally different. And so a couple of things we've done to try to bridge them together. First off, on the cost side, one of the things we realized was that people that are going to run large amounts of on-prem infrastructure increasingly want to do it in the cloud model. And so we introduced a new pricing model that we call ES2, Evergreen Storage Service, which essentially allows you to subscribe to our storage even in your own data center. And so you can have an OPEX experience in the cloud, you can have an OPEX experience on-prem. And when you buy in ES2, those licenses are transferable. So you can start on-prem, move your storage to the cloud with peer, go back and forth, tons of flexibility. From the operational point of view, I think we're trying to get to the same experience as well such that you have a single storage experience from a manageability and automation point of view across both. And I think that last word of automation is key. Because if you look at people who have really invested in cloud, it's all about automation. One of the nice things I think that's made pure so successful in on-prem cloud environments is this combination of simplicity and automation. You can't really automate what isn't simple to begin with. And so we started with simplicity, but as we've added rich APIs, we're really seeing that become the dominant way that people administrate in our storage. And so as we've gone to the cloud, because it's the same software on both sides, literally the same integrations, the same API calls, everything works transparently across both places. You know, that's a great point. We've been reporting on SiliconANGLE on theCUBE for years, Automation's great. If you have to do a couple of manual attaches and automate it, but the value's been shifting. And you guys are in the storage business, you know this data is data. Data is very valuable. You mentioned the car now about to take Uber. Uber's an app. It's got web services in the back end. So when you start thinking about cloud, you think you hear APIs, you hear microservices. As more and more applications are going to need the data, they're going to need to have that in real time, some cases, not near real time, either real time, and they're going to need to have it at the right time. So the role of data becomes important, which makes storage more important. So if you automate the storage, okay, take away the mundane tasks. Now the value shifts to making sure data is being presented properly. This is the renaissance of application development right now, we're seeing this. How do you guys attack that market? How do you guys enable that market? How do you satisfy that market? Because this is where the APIs can be connectors. This is where the data can be valuable, whether it's analytics or an app like Uber that's just slinging APIs together for a service that is now going to go public. Yeah, I think the mindset around data is one of the biggest differences between the old world and the new world. And if you think about the old world of applications, you have monolithic databases that kind of privately own their own data stores. And the whole name of the game was delivering that as reliably as possible, kind of locking it down and making it super reliable. If you look at the idea of a web scale application, the idea of an application is broken up into lots of little microservices. And those microservices somehow have to work together on data. And so what does it mean at the data level? It's not this kind of monolithic database anymore. It's got to be this open, shared environment. And as a result, if you look in the web in Amazon's case, for example, the vast majority of applications are written on S3, object storage that's inherently shared. And so I think one of the bigger, interesting challenges right now is how you get data constructs to actually go both ways. Now, if you want to take an on-prem app that kind of is built around that database, you've got to figure out a way to move it to the cloud and run it reliably. But on the flip side of the coin, if you want to build in web scale tools and then be hybrid and run some of those things on-prem, well, you need an object store on-prem. And most people don't have that. And so this whole kind of compatibility to make hybrid a reality is forcing people on both sides of the wire to understand the other architecture and make sure they're compatible both ways. And to throw more complexity into that equation is that skills gaps. I know cloud needed, but now I'm on-prem and so a different skill gap. You guys had an announcement that's come out. So I want to ask you about your product announcements and your acquisitions. Go back to the past six months. What's the most notable product announcements and acquisitions that you guys have done? And what does that mean for Pure and your customers? Yeah, absolutely. So I'll just kind of walk through it. So the first thing we announced was our new set of cloud data services. And this was in essence bringing our core software that runs on our purity operating environment right into the cloud. And so we call that cloud block store. And again, this is a lot of what I've been talking about, how you can take a tier one block storage application on-prem and seamlessly move it to the cloud. Along that same timeline, we also introduced something called PSO, which is the Pure Service Orchestrator. And this was a tool set that we built specifically for the containers world for Kubernetes, so that basically in a container environment, our storage can be completely automated. It's been really fun watching customers use and just see how different that storage is in a container environment. We look at our call home data within our pure one application and in our traditional on-prem environment, the average array has about one administrative task per day, make a volume, delete something, whatever. If you look in a container environment, that's tens of thousands. And so it's just a much more fluid environment, which there's no way a storage admin is going to do something 10,000 times a day. They've got automation. And that's where automation comes in. But what does that mean, the containerization? That means the clients are using containers to be more flexible. Are they deploying more? What's the insight out of this container trend? You know, I think ultimately it's just a far more fluid environment. It's totally automated and it's built on a world of shared data. And so you need a shared reliable data service that can power these containers. And then back to your original question about kind of product expansion, the next thing that we had announced last year was the acquisition of a company called StoreReduce. And we've subsequently brought out as a product that we call Object Engine. And this is all about a new type of data moving into the cloud, which is backup data and facilitating this backup process. In the past, people moved from tape backup to disk-based backup. And we saw kind of two new inflection points here. Number one, the opportunity to use Flash on-prem so that people have really fast recoveries on-prem because in most environments now disk-based recoveries just aren't fast enough. And then using low-cost object storage in the cloud for retention. So the combination of Flash on-prem and object storage in the cloud can completely replace both disk and tape in the backup process. Okay, so I'm about the competition because you guys came in really with the vision of an all-Flash data center. You now have cloud software that runs on Amazon and others with no hardware, just the block source, which is a great solution. How have the competition fallen behind? Because you guys really kind of catapulted into the lead. Took share certainly from other vendors and my public. Some have predicted that Pure would never make it to escape velocity. Some other pundits and other CEOs of tech companies said that you guys achieved that with their success. Now you guys go to the next level. What is the importance of that ability that you have and what's the inability of the competition? So I like to joke with folks that when we started the company, I think Flash was just an excuse. We just tried to build a better storage company and we went out and I talked to many, many, many customers and I found in general, they didn't just not like their storage products, they didn't like the companies that sold it to them. And so we tried to look at that overall experience and we of course innovated around Flash, used Consumer Flash, bought the price down so you could actually afford to use it with deduplication. But we also just looked at that ownership experience and when I talked to folks in the industry, I think now we might even be better known for our evergreen approach than even for Flash. And it's been neat to watch customers now that even the earliest Pure customers are two or three cycles of refresh and they've seen a dramatic difference in just the storage experience that you can essentially subscribe to and own over time through many generations of technology churn as opposed to that cycle of replacing arrays. Share a story of a customer that's been through those refresh cycles from their first experience to what they're experiencing and what's some of the experiences like and you share some insight. Yeah, so one of the first customers that really turned us on to this at scale was a large telco provider. And they were interesting. They run hundreds of PureOne arrays from competitors and they do a three year cycle. But as they really looked at the cost of that three year cycle, they realized that it was 18 months of usable life in those three years because it took them nine months to get the data on the array and then when they knew the end was coming it took them nine months to get the data off the array. And so per array it was costing them a million dollars just in data migration costs alone. Then you've wasted half of your life of the array. And so add that up over hundreds of arrays in your environment. You can quickly get the math. That's the total cost of ownership. It gets out of control, right? And so as we brought in Evergreen there was just an immediate ROI. I mean it was a cost equation that was on parity with FlashDisk anyway. But if you look at all those operational savings it just, it was completely different. And so I think what we started with Evergreen we realized was much more of a subscription model where people subscribe to a service with us we update and refresh the hardware over time and it just keeps getting better over time. Sounds a lot like the cloud, right? And so we really- And your strategies bring common set of tools in there and bring them again that kind of service. Is that been key? Yeah, yeah I think, you know another thing that we did from day one was we said look we're never going to build a piece of on-prem management software. So our on-prem or our management experience from day one was Pure One which is our SaaS based management platform. You know it started out as a call home application but now as a very full featured SaaS based management experience. And that's also served as well as you go to the cloud because when you want to manage on-prem and the cloud together we're better to do it from then the cloud itself. Tell me about the application environment. You mentioned earlier hybrid multi-clouds here. A lot of pressure on IT to get top line revenue not just cost reduction those are good benefits you mentioned certainly gets their attention but changing the organization's value proposition to their customers is about the experience either app driven or some other tech. This is now an imperative it's happening very fast modernization renaissance and people call it all these things. How are you guys helping that piece of the puzzle? Yeah I mean I think ultimately for most customers as they start to really get in their mindset that technology is their differentiation the speed and agility of their developers becomes key. And so you know a modern CIO is much less about being a cost cutting CIO today and much more about an empowering CIO and how you can actually build the tools and bring them bare for the organization to run faster and a lot of that is about unlocking consumption and so it's been fun to see some of the lessons of the cloud in terms of instant consumption, agility, growth actually come to the mindset of how people think about on-prem as well. And so a lot of what we've done is tried to arm people on-prem with those same capabilities so that they can easily deliver storage as service to their customers so folks can consume via API without having to call somebody to ask for storage so things can take seconds not weeks of procurement, right? And then now as we bridge those models between on-prem and cloud, it becomes a single spot where you can basically have that same experience to request storage wherever it may be in your organization. The infrastructure as code is really just pushing code not from local host or the machine but to cloud or on-prem and just having to kind of trickle all the way through. This is one of the focuses we're hearing in cloud native conversations is words like containers we talked briefly about, you mentioned in the activities high, Kubernetes is really hot right now, service meshes, it's microservices, stateful data, stateless data, these are like really hyped up areas but a lot of traction and force people to take a look at it. How do you guys speak to your customers when they say, hey Kix, we love all the pure stuff, we're on our third iteration or anything about being a customer but I got this looming trend, I got to understand and either operationalize or not around Kubernetes, service mesh, these kinds of cloud native tools. How do you guys talk to that customer? What's the pitch? What's the value proposition? Yeah, I mean I think your new Kubernetes environment is the last place you should consider legacy storage. All joking aside, we've been really, I think positively impressed around how fast the adoption has started around containers in general and Kubernetes, it started out as a developer thing and we first saw it in our environment when we started to build our second product, Pure FlashBlade, four or five years ago, the engineering team started with containers from day one and it was like, oh that's interesting and so we started to see that. It's pretty useful if we use containers and Kubernetes, work are straight, pretty nice and so we just started to see that grow. We also started to see it more within analytics and AI. As we got into AI with ARI and our broader push around going after big data and analytics, those tool chains in particular were very well set up to take advantage of containers because they're much more modern, it's much more about fluidly creating this data pipeline and so it started in these key use cases but I think it's at a point right now where every enterprise is considering it. There's certainly an opportunity in the development environment and despite all of that, the folks who tend to use these containers, they don't think about storage. If they go to the cloud and they start to build applications, they're not thinking many layers down in the organization, what the storage that supports me looks like and so if you look at a storage team's job or an infrastructure team's job, it's to provide that same experience to your container-centric consumers, right? They should just be able to orchestrate and build and then storage should just happen underneath. I totally agree with it. I think that's a success milestone if you can have that conversation and not even be had, you know, you're winning. What they do care about, we're hearing more of what you just mentioned earlier about data pipelining, data they care about because applications will be needing data. Whether it's a retail app or whatever, I might need to have access to multiple data, not some siloed or data warehouse that might have high latency. They need data in the app at the right moment. This has been a key discussion real time. I mean, this is the data, it's been a hard problem. How do you guys look at that solution, an opportunity for your customers? I think one of the insights we had was that fundamentally folks need an infrastructure that can not just run one tool or another tool, but a whole bunch of them. And you look at people building a data pipeline, they're stitching together six, eight, 10 tools that exist today and another 20 that are going to exist tomorrow and that flexibility is key, right? A lot of the original thought in that space was, I'm going to pick the right storage for this piece or the right storage for that piece. But as we introduced our FlashBlade product, we really positioned it as a data hub for these modern applications and each of them requires something a little different, but the flexibility and scale of FlashBlade was able to provide everything those applications needed. We're now seeing another opportunity in that space with DAZ and the traditional architecture. As we came out with NVMeover fabrics within our FlashRay product line, we see this as a way to really take web scale architecture on-prem. You look within Google and Amazon and whatnot, right? They're not using hyperconverge, they're not using DAZ disk inside of the same chassis that happens around applications. They have dedicated infrastructure for storage that's specifically designed for it. They have dedicated servers and they connect them with fast internet, you know, networking on demand. And so we're basically trying to bring that same architecture to the on-prem environment with NVMeover fabric because NVMeover fabric can make local disk feel like, you know, as fast as possible. So this is the shift that's really going on here. This is a complete re-architecture of computing and storage resources. Absolutely, yeah. And I think the thing that's changing it is that need for consolidation. In the early days, I might have said, okay, I'm going to deploy, I don't know, 200 nodes of Hadoop and I'll just design a server for Hadoop with the right amount of disk in it and put them over in those racks and that'll be like this and then I'll design something else for something else, right? Now people are looking for defining a rack that they can print out over and over and over and over again and that rack needs to be flexible enough to deliver the right amount of storage to every application on demand over NVMeover fabric. You know, one trend that I want to get your reaction to is serverless because this kind of points to that value proposition. Functions have been very popular but it's still early days on what functions are. But there's a tail sign a little bit on where this is going to your point around rethinking on-prem, not in the radical wholesale business model change but just more of operating change, how it's deployed and how it works at the cloud because those two things, if working together, make serverless very interesting. Yeah, absolutely. I mean, it's just a further form of abstraction ultimately from the underlying hardware. And so, you know, if you think about functions on demand or that kind of thing, that's absolutely something that just needs a big shared pool of storage and not to have any persistent bindings to anything, you know, to be able to get to the storage and do its task, write what it needs to and get out of the way, right? Well, VP of strategy, a big role. You guys done a good job, so congratulations. Being the number six employee of Pure, how's the journey been? You guys have gone public, still growing. Been around for almost 10 years. You're not a really small little company anymore so you're getting into bigger accounts, growing. How's that journey been for you? Yeah, it's been an amazing ride. That's why I'm still here coming in every day and excited to come to work. I think the thing that we're the proudest of is it still feels like a small company. It still feels like we have as much aggression and much excitement to go out for the market every day as we always have. The orange is very, very strong. But on the flip side, it's now fun that we get to solve customer problems at a scale that we probably couldn't even imagine in the early days. And I'd also say right now, it really feels like there's this next chapter opening up. You know, the first chapter was delivering the all-flash ad center. We're not even done with that yet. But as we bring our software to the cloud and really port it to natively be optimized for each of the clouds, it kind of opens up our engineers to be creative in different ways. Generational shift happening. You're seeing it, you know, again, application modernization, hybrid, multi-cloud are just some key pillars, but there's so much more opportunity to go. I want to get your thoughts. You've had the luxury of being, working under two CEOs that have been very senior veterans, Scott Dietzen and Charlie. What's it like working for both of them? And what's it like with Charlie now? What's the big mandate? What's the hill you guys are trying to climb? Share some of the vision around Charlie's? Well, I'd say the thing that binds both Scott and Charlie together in DNA is that they're fundamentally both innovators. And, you know, if you look at Pure, we're never going to be the low-cost leader. We're never going to be the company that sells you everything. So we have to be the company that's most innovative in the spaces we play in. And so, you know, that's job number one at Pure after reliability. So let's say that's job number two. But that's key. And I think both of our CEOs have shared that common DNA, which is their fundamentally product innovators. And I think the fun thing about working for Charlie is he's really thoughtful about how you run in a company at very large scale, how you manage the customer relationship to never sacrifice that experience because that's been great for Pure. But ultimately, how you also unlock people to run faster in a bigger organization. I had a chance to interview John Chambers who Charlie worked with at Cisco back in the day. He said, one of the key things about a CEO is picking the right wave at the right time. What is that way for Pure? What are you guys riding? It takes advantage of the work you still got to do in the data center on the storage side. What's the big wave? So, you know, look, the first wave was flash. That was a great wave to be on and as I said before, it's not over. But we really see an enormous opportunity where cloud infrastructure mentality comes on prem. And, you know, we think that's going to finally be the thing that gets people out of the mindset of doing things the old way. You know, you fundamentally can take the lessons we learned over here and apply them to the other side of my hybrid cloud. Everybody talks about hybrid cloud and all the thought processes, what happens over in the cloud half of the hybrid. Well, the on-prem half of the hybrid is just as important and getting that to be truly cloud era is a key focus of ours. Yeah, and then again, microservice is only help, accelerate, and you want modern storage to your point to make that work. Absolutely. Okay, thanks for spending the time and sharing the insights. I really appreciate it. It's theCUBE conversation here at Pure Storage Headquarters. We are in the arcade room getting the insights and sharing the data with you. I'm John Furrier. Thanks for watching this CUBE conversation.