 Hi, Cube Nation. Welcome to Cube Conversations. My name is Sam Cahane and I'm coming to you from the Wikibon offices here in Marlboro, Massachusetts. I'm excited to be here as Wikibon senior analyst. Stu Minyan. Stu, thank you for joining us today. Thanks, Sam. Good to chat with you today. Great, great. So today we're going to focus on infrastructure and discuss flash storage and how the market has been evolving. So to start, Stu, what is flash storage? And why is it important? All right. Well, Sam, flash has really been the drumbeat in the storage world for the last few years, really revolutionizing the way we build our infrastructure, as well as even impacting how we're going to build our applications. I love, if you see the little cartoon over my shoulder here, we talked about how really the consumer devices, the smartphones you carry, the tablets, even the laptops that you carry, started using flash storage which is unlike disk, it's non-spinning, it's high performance, it uses low power, had a lot of real advantages when you think about things like battery life and performance that has now come back into the enterprise because flash was actually a technology we used a bunch of years back, 15 years ago when I started working for a big storage company, I saw this big array and it was like, what's that? Oh, that's solid state disk, it's flash. And then really disk was the predominant technology that was used, but flash has really come through and really revolutionized the way we think of storage, think about infrastructure, architecture, and it's been just having a booming effect for the last six years, and we think over the next five to ten years, there will be even more changes going on. So it seems like, and you've seen a lot of changes with flash storage, how has the industry conversation been changing over time? Yeah, great question, Sam. So if we look at flash, when we first started writing about it, people said there's this great little technology, but really where do I use it and how does it fit into what I'm doing? And we said flash, while it is a component, the solutions that we build out of it are many and there's lots of different uses out there. So I think back, when we first did our kind of forecast and definition on flash in 2011, it was you had some flash that was going in the server for specific applications. You had traditional storage arrays that were adding flash, very small percent, 2% flash to really give them greater performance, a little bit more flexibility, and there were these new things called hybrid storage arrays, which mix disk and flash. Fast forward to today, and flash is really pervasive. We have all flash arrays, there's still, of course, hybrid arrays, there's blurring between those. Traditional storage arrays are adding more and more flash, and on the server side, closer to the compute, we're building many more new solutions. So not just a card, but really looking at building out larger solutions closer to the compute. An area that I cover greatly is called hyperconverged, or what we call server-sand at Wikibon allows us to build whole new architectures that take that sole technology of flash, which is increasing in capacity, growing in the utilization and coming down in price, so that it's really changing the way we think about infrastructure, and there's really a solution that we look at today that isn't either leveraging flash, or at least had flashes, some sort of consideration as to how they build things. Great, and I want to focus for a second on the users in this space. So what is the mindset of the user? What are they thinking? Yeah, great point. So we spent the first couple of years really educating users as to what is flash? Do I use it? How do I use it? Oh, it's really expensive. How can I justify it? What does it make sense for? And it's a different conversation today. Today it is an if, it is where, and how much. And we're talking to a lot of users actually, what David Floyer, our co-founder and CTO at Wikibon says is, can you build an all flash data center? And actually, the economics are coming into play that now you can build solutions really built around flash and change that whole environment. So customers are adopting the solutions are growing very fast, both through the startup companies that have filled all of those different segments that I talked about, as well as, you know, the big players in the market, all the traditional storage companies, even the compute vendors are all leveraging flash into what they're doing. So flash is pervasive customers are aware of it. And you know, boy, they're taking advantage of it in a lot of different ways. Great. So, Stu, I want to talk to you a little bit and give you a chance to disclaim any misperceptions in this space. Are there any misperceptions that you see? Yeah. So there's a lot of challenges out there, right? First of all, people think they think flash and they're like, okay, it was kind of high cost. And it kind of was for certain environments. And of course, in the early parts of the technology, the amount of functionality and performance that's growing and the price is going down as we get larger capacities out there in the industry and just more of it out there. So the price economics are changing greatly. One thing David Floyer once again is focused on is the storage efficiencies that I can get with flash, make it so that you can't say, oh, how much do I pay for a gig of disc versus a gig of flash? Because if I can use things like compression and deduplication and other storage efficiency metrics, I can often get, you know, five times more capacity on flash than I could on disc. And we used to have to do things with with disc, things like short stroking and other kind of contortions to try to get performance out of disc that I don't need to flash in many ways can just pull performance out of the equation as a bottleneck and just give you all the performance you want. So I've talked to a number of service providers and large enterprises that said, you know, I used to have to figure out the math as to whether or not I should use flash. But if I just go with mostly flash, I'm deploying all flash arrays or things like that, I actually saves me a lot on my soft costs, my management, my decision points, because I know I've got a baseline in an infrastructure that I can just use. So that's a real thing. What the flash you were looking at two years ago is very different from the flash we were looking at today. And boy, what's exciting from our standpoint is there's a continuation of new technologies and new companies that are coming to bear to bring together new solutions. And the last thing, maybe I don't know if we've got time to talk about it, but it's really helping to drive this next generation of applications. Great, great. And you know, you talk about how two years ago flashes evolved so much. And you know, there's a new world of applications moving forward. How do you see the space changing as time goes on? Yeah, so really, we think this is where we've got the greatest opportunity over the next few years. So of course, when a technology first comes out, it's how do I apply it to what I have. So, you know, six years ago, companies were saying, oh, I did this thing VMware. I'm really starting to deploy it much. I'm virtualizing my applications. I can help get greater performance and take care of what was known as the IO Blender problem, which was, you know, lots of VMs using lots of VMs. You know, go forward today. We're thinking about modern applications. And how do they handle distributed architectures and even more performance? Things like containers and Docker are going to take that IO Blender problem and push it even further. So the architectures will change and the applications are changing to take advantage of these environments. When, you know, 10 years ago, if I built a database, I knew that I had a disk infrastructure and therefore, the way I do my calls and the way I set up my buffers, I know that there's a bottleneck in the system today. If I'm building no SQL databases, if I'm doing analytics and Hadoop, I know I can have much greater performance, whether I'm doing something on-prem or leveraging cloud solutions that are now doing Flash. So we really think that Flash is one of these underlying pieces that's going to help increase what I can do with my infrastructure. And then that is going to drive the next generation of applications which are going to use this new infrastructure. So what we always say, if we spent, you know, five years, we had tremendous growth, the next five years are going to be even more because those fundamental pieces, those applications which we live with for a long time, are going to go through huge changes as Flash just becomes a way of building my infrastructure using that stuff. Great, great. So we only have time for one more question here. So I wanted to ask you, you know, what's next on the agenda for you and your research over Wikibon? All right. So, yeah, I mean, Wikibon, we're really trying to help tease apart. We always want to help the user community understand how to make their choices based on this because, first of all, in IT, unfortunately, everything tends to be additive. So how do I add this into my existing environment? Things like migrations and conversion costs are very hard to go through and usually don't want to do this. So, you know, what we try to tell users is, first of all, try not to get locked into anything. Don't build that next data center because somebody else can actually help you build that better. Don't, you know, try to just hold on to your legacy applications. Let's find ways to modernize what you're doing and leverage these new technologies. We're going out to tons of shows where we're talking to, you know, none of the big players out there, but the smaller startups out there. From a use case standpoint, of course, you know, big data is a big focus area of ours. So the Hadoop marketplace, the NoSQL marketplace, are helping to drive a lot of these conversations and even I've got a strong networking background and Flash is really helping to drive even greater performance. So it puts a strain on the backbone and the connectivity between all my components. So it's always an IT. The bottleneck never goes away. It just gets moved to another part of the infrastructure and I constantly need to just look at holistically what we're doing. And that's what we always advise, is you want to take Flash not as a component or a point solution, but look overall as to how it fits into your business needs, your application requirements and how does it drive value back to what you're doing as a business overall. So it's exciting stuff. Stu Miniman, Wikibon Senior Analyst. Thank you for joining us. Cube Nation, we'll see you next time. Thank you.