 Live from the BuildGram Auditorium in San Francisco, it's theCUBE, covering Pure Storage Accelerate 2018. Brought to you by Pure Storage. Welcome back to theCUBE. We are live at Pure Accelerate 2018 at the BuildGram Civic Auditorium in San Francisco. I'm Lisa Martin, Moonlining is Prince Today, joined by Dave Vellante, Moonlining as the who? Should we call you Roger? Yeah, Roger, Keith, Moonbat. We are, it's a very cool concert venue in case you don't know that. We are joined by a couple of guests, Cube alumni, welcoming them back to theCUBE. Rajiv Vasaredy, the VP of Product Management and Solutions at Pure Storage, and Siva Kumar, the Senior Director of Data Center Solutions at Cisco. Gentlemen, welcome back. Thank you. Thank you. I'm happy to be here. So talk to us about lots of announcements this morning, Cisco and Pure have been partners for a long time. What's the current status of the Cisco Pure partnership? What are some of the things that excite you about where you are in this partnership today? You want to take that, Siva? You want me to take that? I think if you look back at what brought us together, obviously both of us are looking at the market transitions and some of the ways that customers were adopting technologies from our side. The converged infrastructure is truly how the partnership started. We literally saw that the customers wanted simplification, wanted a much more of a cloud-like experience. They wanted to see infrastructure come together in a much more easier fashion, that we bring the IT, make it easier for them. And we started, and of course, the best of breed technology on both sides, being a flash leader from their side, networking and compute leader on our side, we truly felt the partnership brought the best value out of both of us. So it's a journey that started that way and we look back at it now and we say, this is absolutely going great and the best is yet to come. So from my side, basically PureHit started what we now call FlashStack, a converged infrastructure offering, roughly about four years ago. And about two and a half years ago, Cisco started investing a lot in this partnership. We were very thankful to them because they kind of believed in us. We were grown, obviously. But we are not quite as big as we are right now. But they saw the potential early. And so about roughly two and a half years ago, they, I talked about them investing in us. I'm not sure how many people know about what a Cisco-validated design is. It's a pretty exhaustive document. It takes a lot of work on Cisco's side to come up with one of those. So in a nine, and usually a single CVD, it takes about two or three of their TMEs, the highly technical resources, and about roughly three to six months to build those. Per CVD. Per CVD. Wow. And like I said, it's very exhaustive. I mean, you get your bill of materials, your versions, your interoperability, your, you can actually, the commands that you actually use to stand up that infrastructure and the application, so on and so forth. So in a nine month span, they kind of did seven CVDs for us. That was phenomenal. We were very, very thankful that they did that. And over time, that investment paid off. There was a lot of go-to-market investment with Cisco and Pure jointly made. All of those investments paid off really well in terms of the customer adoption and the acquisition. And essentially, we are at a really good point right now. When we came out with our FlashArray X70 last April, Cisco was about the same time they were coming out with the M5 servers. And so they invested again and gave us five more CVDs. And just recently, they've added FlashBlade to that portfolio. And as you know, FlashBlade is a new product offering, not so new, but relatively new product offering from Pure. So we have a new CV that just got released that includes FlashArray and FlashBlade for Oracle. So FlashArray does the online transaction processing. FlashBlade does data warehousing. Obviously, Cisco networking and Cisco servers do everything, OLTP and data warehouse. It's an end-to-end architecture. So that was what Matt Burhead talked about on stage today. We are also excited to announce that we had introduced ARI, AI-ready infrastructure along with NVIDIA at their expo recently. We were excited to say that Cisco is now part of that ARI infrastructure that Matt Burhead talked about on the stage as well. So as you can tell in the two and a half year period we've come a really long way. We have a lot of customer adoption every quarter. We keep adding a ton of customers and we are mutually benefiting from this partnership. So I want to ask you about follow-up on the Oracle solution. Oracle would obviously say, okay, you buy our database, buy our SaaS, buy the RedStack, single-throat to choke, we're going to run better, take advantage of all the hooks we have, you've heard it before. And it's an industry discussion. If customers have it, Oracle comes in hard. So what's the advantage of working with you guys versus going with an all-RedStack? Let's talk about that a little bit. Sure, do you want it? Yeah, I think if you look at the Oracle database as being deployed, this really powers many companies. This is really the IT platform. And one of the things that customers or major customers standardize on is, again, if they have a standardization from an Oracle perspective, they have a standardization from an infrastructure perspective. A database alone is not necessarily easy to put on a different infrastructure, manage them, operate them, go through life cycle. So they look for a architecture. They look for something that's an overall platform for IT. I want to do some virtualization. I want to run some desktop virtualization. I want to do Oracle. I want to do SAP. So the typical IT operates as more of, I want to manage my infrastructure as a whole. I want to manage my database and data as its own island or its own way of looking. So while there are ways to make very appliances behaviors that actually operates one better, the approach we took is truly delivering a architecture for data center. The fact that the network as well as the compute is so programmable and makes it easy to expand really brings a value from a compute perspective. But if you look at Pure again, there's flash arrays truly have world-class performance. So the customer also looks at, while I can get everything from one vendor, am I getting the best of breed? Am I getting the world-class technology from every one of those aspects in perspective? So we certainly think there are a good class of customers who value what we bring to the table and who certainly choose us for what we are. And to add to what Shiva has just said, right? So if you looked at pre-flash, you're mostly right in the sense that, hey, if you built an application, especially with mission-critical application, you wanted it siloed. You didn't want another application jumping in and kind of messing up the performance and response times and all that good stuff, right? And so in those kind of cases, yeah, appliance has made sense. But now when you have all-flash and then you have servers and networking that can actually leverage the performance of all-flash, you don't really have to worry about mixing different applications and messing up performance for one at the expense of the other. It's basically, it's a win-win for the customers to have much more of a consolidated platform for multiple applications as opposed to silos because silos are always hard to manage, right? So yeah. Steve, I want to ask you, Pure has been very bullish really for many years now. Obviously Cisco works with a lot of other vendors. What was it a couple of years ago? Because you talked about this significant resource investment that Cisco has been making for a couple of years in Pure Storage. What is it that makes this, so maybe this flashback, I'm kind of thinking of the three-legged stool that Charlie talked about this morning, but what were some of the things that you guys saw a few years ago, even before Pure was a public company, that really drove Cisco to make such a big investment in this? I think when you look at how Cisco has evolved our data center portfolio, I mean we are in a very significant part of enterprise today powered by Cisco, Cisco networking, Cisco, and then we grew into the compute business. But when you looked at the way we evolved in this compute business, the traditional storage as we know today is something we actually led through a variety of partnerships in the industry. And our approach to the partnership is first of all technology, technology choice was very, very critical that we bring the best of breed for the customers, but also again the customer themselves speaking to us and then our channel partners who are very critical for our enablement of the business is very, very critical. So the way we and when Pure really launched and forade into all flash, and they created this whole notion that storage means flash and that was never the paradigm before. That was a game-changing sort of a model of offering storage, not this capacity, but also flash as my capacity as well as the performance point. We really realized that was going to be a good set of customers will absorb that, some select workloads will absorb that, but as flash in itself evolved to be much more mainstream, every day's data storage can be in a flash media. They realized, customers realized this technology, this partner has something very unique. They've thought about a future that was coming, which we realized was very critical for us. When we evolved network from 10 gig fabric to 40 gig to 100 gig, the workloads that are the slowest part of this any ecosystem is the data movement. So when flash became faster and easier for data to be moved, the fabric became a very critical element for the eventual success of a customer. We realized a partnership with Pure, with the all flash and the faster network and faster compute. We realized there is something unique that we can bring to bear for the customer. So our partnership mindset really said, this is the next big one that we are going to invest time and energy. And so we clearly did that and we continue to do that. I mean, we continue to see huge success in the customer base with the joint solutions. This issue of best of breed versus kind of integrated stacks has been around forever. It's not going to go away. I mean, obviously Cisco in the early days have converged infrastructure, put a lot of emphasis on integrating, on obviously partnerships. Since that time, I don't know what it was, 2009 things have changed a lot. Cloud was barely a thought back then and the cloud has pushed this sort of API economy. Pure talks about platforms and integrating through APIs. How has that changed your ability to integrate best of breed more seamlessly? Actually, I've been working with UCS since it started, right? And perhaps it was a first server system that was built on an API first philosophy. So everything in the Cisco UCS system can be basically anything you can do to it from the GUI or the command line, you could do it through their XML API, right? It's an open API that they provide and they kind of emphasize the openness of it. When they built the initial conversion infrastructure stacks, right? The challenge was the legacy storage ways didn't really have the same API first programmability mentality, right? If you had to do an operation, you had a bunch of a ton of CLI commands that you had to go through to get to one operation, right? So Pure, having the advantage of being built from scratch when APIs are what people want to work with, does everything through REST APIs. All function features, right? So the huge advantage we have is that with both Pure, Pure actually unlocks the potential that UCS always had to actually be a programmable infrastructure. That was somewhat held back, I don't know if I agree or not, but I will say it, that kind of was held back by legacy hardware that didn't have REST-based APIs or XML or whatever. So for example, they have Python and PowerShell-based toolkits based on their XML APIs that they built around that. We have Python PowerShell toolkits that we built around our own REST APIs. We have puppet integration, salt and all the other stuff that you saw on the stage today and they have the same things. So if you're a customer and you've standardized, if you've built your automation around any of these things, right? If you have the end to an infrastructure that is completely programmable, the cloud paradigms that you're talking about is mainly because of programmability, right? That people like that stuff. So we offer something very similar, the joint value proposition. You're bringing that DevOps kind of infrastructure as code mentality to systems design and architecture. And it does allow you to bring the cloud operating model to your business on- An aspect of the cloud operating model, right? There's multiple different things that people think about. Yeah, maybe not every single feature, but the ones that are necessary to be cloud-like on- Yeah, absolutely. That's kind of what the goal is. Let's talk about some customer examples. I think Domino's was on stage last year and they were mentioned again this morning about how they're leveraging AI. Are they a customer of FlashJoc? Because that may be something you can kind of dig into. Let's see how the companies that are using this are really benefiting at the business level with this technology. No, I think, absolutely. Domino's is one of our top examples of a FlashJoc customer. They obviously took a journey to actually modernize, consolidate many applications. In fact, interestingly, if you look at many of the customer journeys, the place where we find it much, much more valuable in this space is the customer has got a variety of workloads and he's also looking to say, I need to be cloud-ready. I need to have a cloud-like concept, whether I have a hybrid cloud strategy today or it'll be tomorrow, I need to be ready to be able to consume on-prem and cloud. And the customer also has the mindset that while I certainly will keep my traditional applications such as Oracle and others, I also have a very strong interest in the new and modern workloads, whether it is analytics or whether it is even things like containers, microservices and things like that, which brings agility. So while they think that I need to have a variety of things going, then they start asking the question, how can I standardize on a platform, on an architecture, on something that I can reuse, repeat, and simplify IT? That's by far, it may sound like you've got everything kind of thing, but that is by far the single biggest strength of the architecture, that we are versatile, we are multi-workload, and when you really build and deploy and manage, everything from an architecture, from a platform perspective, looks the same. So they only worry about the applications they're bringing on board and worry about managing the lifecycle of the apps. And so a variety of customers, so what has happened because of that is, we started with commercial or mid-sized customers to larger commercial, but now we are much more in enterprise, large, many large IT shops are starting to standardize on Flash Tag, and many of our customers are really measured by the number of repeat purchases they will come back and buy, because once they like and they bought, they really love it and they come back and buy a lot more. And this is the phase where, it gets very exciting for all of us that these customers come back and tell us what they want. Whether we build automation or whether we build management architecture, our customer speaks to us and says, you guys better get together and do this. That's where we want to see our partners come to us and say, we love this architecture, but we want these features in there. So our feedback and our evolution really continues to be a journey driven by the demand in the market, driven by the customers who we have. And that's hugely successful when you are building and launching something into the marketplace. Your best reward is when customer treats you like that. So to basically do a dovetail into what Shiva was talking about in terms of customers, so he brought up a very valid point. So what customers are really looking for is an entire stack, an infrastructure that is near invisible. It's programmable, right? And you can kind of cookie cutter that as you scale. So we have an example of that. I'm not going to use the name of the customer because I'm sure they're going to be okay with it, but I just don't want to do it without asking the permission. It's a healthcare service provider that has basically literally dozens of these flash tags that they've standardized on. Basically, they have vertical applications, but they also offer VM as a service. So they have cookie cutter this with full automation integration. They roll these out in a very standard way because of a lot of automation that they've done. And they love the flash tag just because of the programmability and everything else that Shiva was talking about. With new workloads coming on, do you see any architectural limitations, what am I saying, new workloads? Data-driven, machine intelligence, AI workloads. Do you see any architectural limitations to scale and how do you see that being addressed in the near term? Yeah, that's actually a really good question. So basically, let's start with the... So if you look at bare metal VMs and containers, that is one vector. In that vector, we're good because we support bare metal and so does the entire stack. When I say we, I'm talking about the entire flash tag, servers and storage and network VMs, and then also containers because most of the containers in the early days were ephemeral, right? Then persistent storage started happening and a lot of the containers were deployed in the public cloud. Now, we are getting to a point where customers are basically experimenting with large enterprises with containers on-prem. And so the persistent storage that connects to containers is kind of nascent, but it's picking up. So there's Kubernetes and Docker are the primary components in there, right? And the Docker, we already have Docker native volume plugins and Cisco has done a lot of work with Docker for the networking and server pieces. And Kubernetes has flex volumes and we have Kubernetes flex volume integration and Cisco works really well with Kubernetes. So there are no issues in that vector. Now, if you're talking about machine learning and artificial intelligence, right? So it depends. So for example, Cisco's servers today are primarily driven by Intel-based CPUs, right? And if you look at the NVIDIA DGXes, these are mostly GPUs. Cisco has a great relationship with NVIDIA and I will let Shiva speak to the machine learning artificial intelligence pieces of it, but the networking piece for sure, we've already announced today that we're working with Cisco in our ARI stack, right? So... Yeah, no, I think the next generation workloads or any newer workloads always comes with a different set of, some are just software level workloads. See, typically software type of innovation, we, given the platform architecture is more built with programmability and flexibility, adopting our platforms to a newer software paradigms such as container microservices, we certainly can extend the architecture to be able to do that and we've done that several times. So that's a good area that covers, but when there are new hardware innovations that comes with either it's interconnect technologies or whether it is new types of flash models or machine learning GPU style models, what we look at from a platform perspective is what can we bring from an integrated perspective that of course allows IT to take advantage of the new technology, but maintain the operational and an IT cost of doing business to be the same. That's where our biggest strength is, of course NVIDIA innovates on the GPU vector, but IT doesn't just do GPUs, they have to integrate into a data center, flow the data into the GPU, run compute along that, run applications to really get most out of this information and then of course process it for any kind of real time or any decision making for that matter. Now you're really talking about bringing it in house and integrating into the data center. Anytime you start in that conversation that's really where we are, I mean we welcome more innovation, but we know when you get into that space we certainly shine quite well. It's secure, it's protected, it's move it, it's all kinds of things. So we love these innovations but then our charter and what we are doing is all in making this experience of whatever the new be as seamless as possible for IT to take advantage of that. Wow guys, you shared a wealth of information with us. We thank you so much for talking about the Cisco Pure Partnership, what you guys have done with Flashdack, how you're helping to customers from pizza delivery dominoes to healthcare services to really modernize their infrastructures. Thanks for your time. Thank you very much. For Dave Vellante, I am Lisa Martin. You're watching theCUBE live from Pure Accelerate 2018. Stick around, we'll be right back.