 Live from San Francisco, it's theCUBE. Covering Red Hat Summit 2016. Brought to you by Red Hat. Now, here are your hosts, Stu Miniman and Brian Gracely. Welcome back, happy to welcome back to the program. Ranga Rangachari, sorry, who's Vice President General Manager for Storage with Red Hat. Ranga, thanks for joining us. Thank you, thanks a lot. All right, so it's been container week here. Red Hat Summit, a lot of the announcements. Give us the, you know, how containers and storage, especially from a SAP standpoint, go together. Absolutely, and, you know, as customers start on their container journey, as you pointed out, there's just been a phenomenal positive feedback on the container announcements over the last few days. But fundamentally what this means is developers, when they start to get into container and container application platforms, they want storage to be as invisible as possible. They don't want to be in the business of calling up somebody and say, hey, give me storage for this application and that application. So what we've announced as part of the container application platform is what we call container native storage. So this is where containers and storage, the server and the storage are hyperconvergent in an environment where you can actually build, scale your application. It's all orchestrated through Kubernetes. So the biggest value for the DevOps and the developers is, one, storage is invisible from that sense. And two, it scales as their deployment scale. So it's totally a great opportunity for the developers to make their container journey successful. And the way it's actually manifests itself is we came up with a new release of Red Hat Luster Storage 3.1.3 a couple of days ago. And that's the foundational piece of our container native storage for our container application platform. OK, so does that include persistence for storage in containers? Absolutely. So there are two things. And this is when we talk about, that's a great point. When we talk about container native storage, it's persistent, especially when you're building stateful applications. The need for persistence storage is very, very key. Case in point would be one of the nodes might either be retired or might die for whatever reason. How do you make sure your storage doesn't die alongside with that? So the persistency and the scalability of storage is something that when we talk to our customers the last couple of days, they just think that's a very, very integral part. Because we contrast that with how it's been done before, is you need a separate set of storage arrays that provide storage to your application. That's very cumbersome, very friction-prone. So you've taken the friction and the ease of deployment because developers really look for time to value. This really helps them take it to the next level. Yeah, I guess it's quite different from what we saw with virtualization. Server virtualization, while it broke storage, it was just, I couldn't see at the VM level. From here, containers have some different characteristics. As you said, sometimes I might just have storage separate and I can be much more short-term with my containers. What do you see from the application side and how fast are we going to mature that solution for storage? Well, I mean, the early indications, and I think Paul yesterday talked about the customers, the early indications are very positive. After all, being two or three days since we announced this, but just based on the customer conversations that we've had here at Summit over the last two days, the interest level and the level to start to deploy this in an early proof-of-concept phase is here and it's here now. So all indications of this is really going to help change the DevOps and help people actually use OpenShift under container application platforms. Yeah, there was a great line I saw the other day that said there's really no such thing as stateless applications. You just punted the problem to somebody else. You look at some of the customers that were referenced here, Amadeus won the Innovation Award, GenTech won the innovation. Amadeus is doing transactions. That's a stateful thing. GenTech's doing transactions for transmission. I think people sort of forget they hear about 12 factor and they think, oh, that's containers. There's a transaction going on somewhere you want to maintain data. How much are people starting to understand that that's a critical role in these new types of applications? If you asked this question six months ago, I think it was a lot of what I would call evangelism. We had to highlight the problem of how it's very critical that you build out true stateful applications. But now I think the level of familiarity and the level of actually understanding is a night and day in terms of what things are. So it's about, in a way I think about this is it's classic technology adoption with early adopters and I think we are kind of getting past that stage where it's becoming more and more almost de facto if this is the way you got to build your apps. And storage has to be, you know, you need the other hand to clap if you will, right? So you need to get the applications rolled out. Storage has to be part of that. Yeah, we've seen so much change. I mean, Paul talked about it. Jim talked about software defined storage. So moving away from arrays, we're seeing containers play a role. Who's using this? Is it still somebody who would identify as a storage admin or what's the role that needs to know about container storage and stateful storage for containers? That's a great question. And I think 12 to 18 months ago, the conversations around software defined storage was primarily I would say the storage admin, right? But now the shift has happened. I think the shift primarily is because of cloud, right? And it's no longer a storage conversation, it's a cloud conversation. Now, yeah, if you double click into it, they might be somebody in the storage as part of the cloud team who's responsible for storage. But it is about how do you build a scale out infrastructure? Whether it's scale out compute or scale out storage or scale out networking, how do you have software defined? How do you build that out? And that's one conversation instead of conversation. The other sort of conversations, especially with containers, it's becoming more of a DevOps conversation. So it's not that the storage, the traditional storage buyers are less relevant. It's more a question of they're providing a supporting role to the business leaders and to the cloud architects to make sure that architecture is sound and it can scale. Can you speak to just from the software to sign storage, we look at what use cases, what applications, you know, where does it fit? And maybe where are the areas that you say, you know, it's not quite there yet, but it's on the road now. Yeah, so I think one of the primary use cases for Red Hat Cef storage is around OpenStack, right? If you look at, and I think you guys have been in a lot of these OpenStack shows, last OpenStack survey, I think that it's 62% of customers are running OpenStack and actually using Cef as a storage substrate. So it's becoming a majority or a de facto, if you will, as customers are deploying. So that's how do you store your VM images and fire them up and not in tens or hundreds but thousands of those things. So that's a use case that shines really well. Now, part of the thing that we announced last week is object storage. And this is where as the type of data, you know, audio, video, rich media, all those things start to becoming, we are not talking about few hundred terabytes. These are petabytes, multi-petabytes, 50, 60 petabyte use cases where customers are just grappling with how do I manage this? And Cef is a perfect answer for that. So those are the type of uses. On the cluster side, it's still a lot of use cases around what I would call secondary storage and archiving, cheap and deep type of things. And now with the container offering that we have, we will see more of those container native storage becoming a fabulous use case for the cluster solution. All right. And how do these storage solutions fit in with the various cloud offerings, especially I'm curious, you know, public cloud. Yeah. We have a lot of exposure here. They actually fit in very well. So just go ahead, stepping back a little bit. When we look at the landscape, right, you got physical environments, virtual environments, private cloud, public cloud, and now containers, right? And the Red Hat Storage portfolio fits perfectly well with each of those things. So six months ago when we announced our relationship with Microsoft, Red Hat Gluster Storage is now available in Azure, right? Three months ago, we announced where Google is part of the Google Cloud Platform. Red Hat Gluster Storage is available. Customers can buy through a cloud access program, buy our storage offering and move it to a public cloud. And the cool thing is because of the way, in this specific case, the way Gluster is designed, you don't need any application rewrites. So you can run your application on-prem and for whatever reason you want to move into one of these public clouds, whether it's Google or Amazon or Microsoft, nothing changes. Say just it's an army or it's whatever format, from an application because of the POSIX compliance, the application doesn't have to be rewritten. So that's why the application developers really start to value that. Yeah. How big a role is Flash playing in SEP and high-performance applications? That's a great question. And that's the other macro trend that we are seeing more and more pronounced over the last six to nine months, where Flash or SSDs are almost becoming the latter-day HDDs, right? And so far, I would say that a lot of the SEP and the Gluster use cases were primarily on what I would call capacity optimized workloads. Now with Flash, it's becoming more and more performance optimized workloads. So a few months ago, we announced a relationship with SandDisk, where SandDisk, InfiniFlash is now supported with SEP. And those are ideally, so what we've done, we've taken that a step further, what we've actually done is created reference architectures for our customers. Because even though we talk about software-defined storage, ultimately customers say, okay, how does it relate to the piece of hardware they want to run this on? And we've taken some of the, a lot of the friction out of the system by providing reference architectures where the customers say, great, now I know how I can marry an InfiniFlash with SEP storage, put it together so that I can get the best performance out of it. Yeah. Can you speak to how you see that go to market for storage changing? I think historically, I mean, storage has been a box. And today, if I buy SaaS, I, a lot of times, don't think about it. Public cloud, I'm consuming storage without thinking about it. How do you see the go to market changing? So one of the, I guess, the beautiful things about software-defined storage, it gives you multiple options on routes to market. If this was in a hardware, the traditional monolithic storage mainframes, as we call it, your routes to market are very limited, right? You got this piece of hardware, you got this piece of software, you got the only way you can get it is combine them together, push it through your channels of distribution. But for us, given the flexibility that our solutions offer, we have a wide range of go to market options. What I would call the direct go directly to the telcos, to the enterprises, sell this. But more importantly, there is embedded options. There is this public cloud options. Then there is things like classic bars who want to build solutions into the place. So one of the things that we are focused on right now, predominantly our solutions in the marketplace today have been what I would call direct catch, right? But now we are opening it up to certified channel partners around the country, around the world who can really start to build their practices on software-defined storage. They're very, very excited because they're customers and are asking them, what's your SDS offering in the marketplace? And a flip side of that, where do you see the role? You talked about telcos. I know when we were at OpenStack Summit a few weeks ago, Verizon was on, Verizon, not only OpenStack customer, but SAP customers as well. Where do you see that playing in terms of giving customers choice and adoption in the marketplace? So from a telco perspective, if you view the telcos as our customers, I think the fact that Red Hat has a portfolio of storage offering gives them the ability to pick and choose depending on what workloads they want to address to their customers. So the typical telco layout, if you will, is very much a small, medium-large, where they have large data centers, they're talking about petabytes, and they have these reasonable points of presence. With there, it's very much of a hyperconverged environment. They don't have the footprint. They want to say, like, I want to pack in as much as computer storage as possible in a closet out there. And to their customers, essentially, it's kind of invisible. They just go in and say, okay, I need to consume storage. Here it is. So Ranga, I want to give you the last word. I know your team's been doing a lot all over the show. For those that haven't been able to come to the Summit itself, what would some key take-aways or things you'd want them to look at that they might have missed? Yeah, so let me put this in context of what the customers that I've been meeting with today have been telling me. And there's a common theme that's emerging, which is, I cannot run two days and tomorrow's workloads on yesterday's infrastructure. And that's really starting to change the conversation. All right, great. That's a great message. Ranga Rangachari, really appreciate you giving us the update on all things storage here at Red Hat Summit. We'll be right back with more coverage from San Francisco at the Moscone Center West. You're watching theCUBE.