Narrator: From Austin, Texas, it's theCUBE covering Pure Storage Accelerate 2019. Brought to you by Pure Storage. Hi, Lisa Martin with theCUBE. Dave Vellante is my co-host. We are at Pure Accelerate 2019 in Austin, Texas. One of our Cube alumni is back with us. We have Rob Lee, the VP and chief architect at Pure Storage. Rob, welcome back. Thanks for having me back. We're glad you have a voice. We know how challenging these events are with about 3000 partners, customers, press, everybody wanting to talk to one of the men that was on the keynote stage. Yesterday, four announcements came out. Really enjoyed yesterday's keynote, but let's talk about one of those announcements in particular: Pure's bridge to the hybrid cloud. Absolutely, yeah, I think it's been a really exciting conference for us so far. Like you said, a lot of payload coming out. As far as the building the bridge to the hybrid cloud, this has been, I would say, a long time coming. We've been working down this path for a couple years. We started by bringing some of the cloud-like capabilities that customers really wanted and were able to achieve into the cloud back into the data center. So you saw us do this in terms of making our on prem products easier to manage, easier to use, easier to automate. But working with customers over the last couple of years, we realized that, as the cloud hype kind of subsided and people were taking a more measured view of where the cloud fits into their strategies, what tools it brings, we realized that we could add value in the public cloud environment the same types of enterprise capabilities, the same type of features, rich data services, feature sets, things like that that we do on premise in the cloud. And so what we're looking to choose is actually quite simple. We want to give customers the choice. Whether customers want to run on premise or in the cloud, that's just a choice of, we wanted to make an environmental choice. We don't want to put customers in a position where they have to make that choice and feel trapped in one location or another because of lack of features, lack of capabilities, or economics, and so the way that we do that is by building the same types of capabilities that we do on prem in the cloud, giving customers the freedom and flexibility to be agile. But, you know, you mentioned economics and you were talking from a customer standpoint. I want to flip it from a technology supplier's standpoint. The economics of a vendor who traditionally sells on prem, you would think, would be better than one in the cloud because you got, you're paying Amazon for all their services, or I guess the customer's paying for it. But you saw your way through that. A lot of companies would be defensive and I wonder if you could add any comment to that. Yeah, no, I mean, so look, I think... The hardware's only one piece of it. At the end of the day, even our products on prem are really priced for value right there. We're delivering value to customers in our capabilities, our ease of use, our simplicity, the types of applications and workflows we enable. And basically everything I just said is pretty much driven by software features. By bringing those same capabilities into the cloud, naturally, most of that work is really in software. And then, as far as comparing the economics directly on prem versus cloud, it's really no secret as the industry has gotten more understanding that the cloud isn't the low-cost option in a lot of use cases. So, rather than comparing apples to apples on prem versus cloud, either on performance or economics, our goal is really to build the best products in either environment. So, if a customer wants to run on prem, you want to build the best darn products in that environment. If a customer wants to run in the public cloud, we want to build the best darn product for them in that environment. And increasingly, as customers want to use both environments hand in hand, we want to build the right capabilities to allow them to seamlessly do that. Well, I think it makes sense because, as you know, we're talking to some customers last night. We ask them what they have in your data center. They got a lot of stuff in their data center to the extent that a company like Pure can say, okay, you've got simple, fast, et cetera on prem and that we've now extended that to the cloud. Your choice. They're going to spend more with you than they are with the guys that fight that. Yeah, absolutely, and I think, if you look at our approach in how we build the products and how we're taking them to market, we've taken a very different approach than some of the competitive set. In some ways, we've really just extended the same way that we think about innovation and product engineering from our existing on prem portfolio into the cloud, which is, we look for hard problems to solve. We take the hard road. We build differentiated products even if it takes us a little bit longer. You can see that in the product offerings. We've really focused on enabling tier one mission critical applications. If you look at the competitive set, they haven't started there. The reason why we did that is we knew that we had customers telling us. If you're a customer and you want to use the cloud and you want to think about the cloud as a DR site, well, when something goes wrong and you have to fail over the DR site, you need to be sure that it works exactly the same way there as it did on prem. That's everything from data services, data path features, to all the workflows and orchestration that go around it because, when your primary site goes down, it's not the time when you want to be discovering that, oh, there's a footnote on that feature and that's not supported in the cloud version, that sort of thing. And so, like I said... The focus that we've put on the product development we've done towards cloud block stores really been arounds creating the same level of enterprise grade features and enabling those applications in the cloud as we do on prem. You know, we don't make the Amazon storage. We make the Amazon storage better. What's that commercial? >> (laughs) That's essentially what you've done. That's essentially what we've done, yeah. And the great thing about that is that we've done it in close partnership with Amazon. We had Amazon onstage yesterday and they were talking a little bit about that partnership process and, ultimately, I think why that partnership has been so successful is we're both ultimately driven by the same thing which is customer success. In the early days of working with Amazon, as we started coming up with the concept of Cloud Block Store and consulting them on, "we're thinking about building it this way, "what do you think? "What services should we leverage in AWS "to make this happen?" It became pretty clear to them that we were setting out to build a differentiated product and not just tick off check boxes. And that's when their eyes really opened. Yeah, I'll bet. >> Hey, we really would like you to do a differentiated product here, so-- Hey, if this takes off, we're going to sell a lot of EC2, S3 (chuckles). What are some of the things... Sorry, Dave, but you've been with Pure about six years. What are some of the things that have surprised you pleasantly that the customers have catalyzed from an architecture perspective? The customer feedback, coming back to your team, and the guys and girls engineering their product. Customers are demanding a certain thing that maybe wasn't something that was an internal idea but really was catalyzed by customers. Anything that just really, you think, is just very cool or very surprising? Yeah, no, I mean, I think a couple of things. I think, personally, one of the things that surprised me was when I joined Pure in 2013, we're all about simplicity. You talked to Koz who I think you had on the show earlier. In the early days, he'd tell you our differentiator is going to be simplicity, and I got to say, when I first ran the company, I was a little skeptical. I was like, all right, I get it. Simplicity is a thing, is it really a differentiator? I very quickly was surprised based on customer feedback that, no, it really is very, very meaningful. And that's something that we take all the way through engineering, everything down to how we design features and put them in the user interfaces. If there's an engineer that wants to put a configuration hook or a knob or an option in the user interface, we kind of stop and say, well, gee, how would you document that? How would you suggest a user make a decision to set that value? They'll describe and say, okay, well, gee, we can make that decision, can't we? Why don't we just make it simpler? And so that's been the biggest surprise. I think, from a customer-catalyzed point of view, what I'd say is we've been really surprised at a lot of the use cases that the FlashBlade product has been put into play for. And I think AI was one of them. When we first set out, we had really targeted FlashBlade at addressing a segment of the commercial-age PC chip design, hardware design, software development market. And it's actually a set of customers. Very large web property customer that came to us with an AI use case, they said, "hey, we've got a ton of data, "video, images, text postings, "and we want to do a lot of analysis of this. "We want to do facial recognition, "we want to do content and sensitive analysis. "We've got the GPU's, we think you guys "have the right storage product for that." And that's really taken off, and that was very much a customer-driven area. You talked a little bit about that with Nvidia yesterday about some of the customer-catalyzed innovation where AI is concerned. Absolutely. What do you see as the critical technical skills that Pure needs in the next decade? I mean, correct me if I'm wrong, but you have a networking background, internal networking, I guess, of, you got guys from Veritas, right? Obviously strong software file system. What do you see is the critical skill? Yeah, that's a good question. We have a very diverse team. We, in engineering, typically hire and look for people with strong systems, backgrounds that are willing to learn and want to solve our problems. We typically haven't hired very specific domain areas. Myself, my background is in language runtimes and compilers. Oh (chuckles). >> I distributed systems, so a bit all over the map. You know what I'd say is that the first phase of Pure, the first decade, was really about reinventing the storage experience, and for me, I look at it as taking lessons from the consumer experience, bringing them into the storage and enterprise world. The iPhone's example, that's used a lot. There's a couple other examples you can think of. I think the next phase of what we're trying to do and you heard Charlie talk about this onstage with the modern data experience is take some lessons from the cloud experience and bring them into the enterprise. So the first phase was about consumer simplicity, it's for a human. Think the next phase is really about bringing in some more of the cloud experience for enabling automation and devops and management orchestration. So what kind of work? I mean a lot of work to do to get... We envisioned this massively scalable distributed system where you have that cloud experience no matter where your data lives. That's not there today. And you don't want to ship your data around. There's going to be too much data, so you want to ship metadata and have the intelligence to bring the compute to that data. What do you got to do? What's the work that you have to do to actually make that "seamless"? There's that overused word again. It's not seamless today. Yeah, so I think there's a lot of angles to it. And we're going to work our way there to your point. It's not there today. But you're starting to see us lay the groundwork with all the announcements that came out today. Under the umbrella of, hey, we want to end up creating a more portable, more seamless, more agile experience for our customers, you can see where, as we bring more storage medias into play, different classes of service, different balances of performance and cost, bringing those together in a way so that an application can use them in the right combinations. Bringing AI into play to help customers do that seamlessly and transparently is a big part of it. You can see multiple location kind of agility that we're bringing into play with Cloud Block Store. Enable the CloudSnap and snapshot mobility, things like that. And then I think, as we move beyond the block world and we look at what we can enable with applications that sit on top of file and object protocols, there's a lot of greenfield there. So we think object storage is very attractive and we're starting to see that, as the application renders, as the applications that sit on top of a storage layer are really embracing object storage as the cloud-native storage interface, if you will, that's creating a lot of... A lot of ways to share data. We're starting to see it even within the data center where multiple applications now are able to share data because object storage is being used. And so, like I said, there's a lot of angles to this. There's bringing multiple discreet arrays together under the same management plane. There's bringing multiple different types of storage media a little bit closer together from a seamless application mobility perspective. There's bringing multiple locations: data centers, clouds together from a migration, a DR perspective. And then there's bringing a global namespace type of capability to the table, and so it's a long journey, but we think it's the right one, and what we ultimately want to do is have customers be able to think about, be able to provision, be able to manage to not just an array, but really more of like an AZ. I want a pool, I want it to be about the A fast, but I'm willing to pay about your A much for it and I need these types of data protection policies for it. Please make it happen. And anywhere, so you see it as technically feasible to be able to run any app, any workload on any cloud or on prem without having to recompile the application, make changes to the application. That's sort of really kind of meant by seamless that you see that as technically feasible in the next, we'll call, five to 10 years, I'll give you. I think it'll take a long time to get there and I think it'll depend on the application. I think there are going to be some combinations that, if you have a high frequency, low latency, trading database, there's physical limitations. You're not going to run the application here and put the storage in the cloud. But if we step back from it, the concepts, I mean I think that a lot of things are becoming possible to make this happen. Faster networking is everywhere, it's getting faster. Application architectures are making it more feasible. The media costs and what we're able to drive out of the media are bringing a lot more of those workloads to Flash. Dave: AI, you mentioned. >> AI is coming into play, so like I said, it's going to depend on the application, but I think we're entering a phase where the modern software developer doesn't want to have to think too hard about where, physically, what six sides of sheet metal is my data sitting on. They want to think about what do I need from it in terms of capacity, what I would need from it in terms of performance, what I would need from it in terms of data service capabilities, and I need to be able to control that elastically. I need to be able to control that through my application, through software, and that's kind of what we're building towards. Last question, Rob, as we wrap up here. Feedback that you've heard the last day and a half on some of the news that came out yesterday from customers, analysts, partners. Yeah, I'd say, if I were to net it out, I think the one piece of feedback we've gotten is "Wow, you guys have a lot of stuff "and it's really nice to see you guys "talking about stuff that's available today." A lot of GA's on that screen yesterday. A lot of GA's on that screen and I think I had an analyst say to me, "It's really refreshing to see you guys "take A, both the viewpoint of the customer, what you're delivering the customer, what you're enabling, and then B, I go to a lot of tech conferences and I hear a lot about way off in the future and vision and the feedback we got was "You guys had a really good balance "of reality today, what you're helping customers with today, "what's available today to do that, "and enough of the 'hey, and here's where we're headed.'" We actually heard the same thing, so good stuff, Rob. Well, congrats on the >> Thank you. tech anniversary and we appreciate you joining us on theCUBE. We're looking forward to next year already in whatever city you're going to take us to. Me too, me too, thanks a lot. All right. For Dave Vellante, I'm Lisa Martin. You're watching theCUBE. Thanks for watching. (upbeat electronic music)