 Live from Mountain View, California, it's The Cube, at OpenStack Silicon Valley, brought to you by headline sponsor, Mirantis. Here are your hosts, John Furrier and Jeff Frick. Okay, welcome back here to the OpenStack Silicon Valley live broadcast for The Cube, our flagship programming off of the events. Extract the ceiling from the noise. I'm John Furrier, the founder of SiliconANGLE, showing my co-host, Jeff Frick, here in Palo Alto for The Cube office. Our next guest, John Dickerson, with Swift Stack and the OpenStack Swift Project technical lead, welcome to The Cube. Thank you very much for having me. So obviously, storage is the epicenter. We cover all the storage shows, EMC, VMware, all that's going on in flash. Storage is where the action is, big data, big part of the cloud, mobile, social, data exhaust, Internet of Things, just the list goes on and on. So you're in the battleground. Okay, so talk about what's going on with Swift and some of the design decisions you're making, what new projects are coming out, some highlights, give us the quick update. Sure, in Swift we've been doing some great things recently. We just recently completed a year-long development effort inside of the community, led by people from Swift Stack and Red Hat and Box and Intel to put together a feature we're calling Storage Policies. And Storage Policies let you choose what subset of hardware you store your data on and then how you store it across that subset of data. So you can set up geographic localization, you can set up different performance tiers all inside of the same managed, the same cluster. You could even do things that are along the lines of, I want to choose different replication models. And we're extending this right now by working on implementing erasure codes as a storage policy inside of Swift. So you can have your global Swift cluster that is, here's data that's stored in one local place, here's data that's stored in another place, here's data that's distributed throughout it and different data can be replicated versus erasure coded versus reduced redundancy replication just all very specifically depending on your use case. And I think that's the exciting thing about it is that it really helps, it gives applications the actual, the power and the flexibility that they need so the applications can continue providing their value and that they can become more effective and more efficient in how they're consuming object storage. So what's the big change between last Atlanta and now? Obviously Atlanta, big part of the conversation, certainly Docker is being talked about as a container not necessarily related to what your project is, but we are talking about application developers who want infrastructure as code. That's the DevOps ethos. But with Hybrid Cloud, IT is not so much DevOps oriented but they just want the same benefits. So I think that there's been a lot of really great development in that sort of thing. Looking at deployment models but very specifically looking at how do you provide value to the end user and that comes from those applications. So my view is that we very much need to be focused on giving the tools and flexibility to people who are application developers so that they can then develop the, they can focus on providing their value add and the storage system like Swift can offload the hard problems of storage so that people can actually focus on solving the real world use cases that they have. So I don't know if you have any tangential data from the customers but talk a lot about Flash and the growth of Flash and the evolution of Flash and moving from really kind of high value low latency applications and more general purpose applications and really Flash kind of native applications. Are you seeing a lot of that or is that something you guys even, is that one step above where you're playing? No, it's not. It's about the adoption of the Flash. It's not at all where we are, it's not at all above where we are right now, but it's definitely something that is coming on the horizon. Flash is incredibly exciting and offers a lot of really new possibilities, a whole lot more than just it's faster. But you actually have new usage models that can really take advantage of all the native characteristics of that storage in Silicon. But, so we're definitely keeping our eye on that. So we're seeing it a little bit now with looking at high performance tiers and things like that but it's not quite yet at the cost point where it's competitive for the large, I mean, we're talking petabytes and petabytes of storage sort of thing. In that use case, it's not a huge market driver right now but it's definitely something we're looking at on the horizon. Yeah, and clearly if it's software defined, they can tweak that in real time based on application demands and or specific use cases or? The whole value of OpenStack Swift is that with the storage system you've decoupled the data from the actual infrastructure that it sits on, which means that you can have your system, you don't have to worry about what data is it stored on, what happens when that fails, what life cycle is that data in and how does the deployer rotate it out from behind? Which means that your clusters can continue to get better over time as you get newer hardware and move to things like Flash and stuff like that. So yes, I think it's absolutely important and critical. And what are people more excited about the cost savings or people more excited just about higher utilization or people more excited about really the dynamic structure that enables them to do things that they couldn't do before with that decoupling? I think it's a combination of all of those, but the biggest, probably the thing that first attracts people to that is the fact that you don't have to deal with the management of your storage infrastructure. You can treat it as a consumable utility. That's the promise of the cloud, right? And so if you've got an application that is, well, maybe you look at guys like mobile phone apps and things like this, they're being bought for hundreds of millions or billions of dollars, but they don't have to worry about how do we deal with the concurrent data usage on across millions of devices and scaling that out. They can simply just treat it as a resource by using object storage. What's next? What are you excited about? What are we going to talk about a year from now? Well, I think that in a year from now, inside of OpenStack Swift, you'll see Erasure Codes implemented. You'll also see other capabilities built on top of storage policies, and we're also working on ensuring that we can really build out some of the efficiencies of the system so that you can get denser deployments, you can get better cost savings overall. Overall, my view and my point, my goal as the Project Technical Lead for Swift is to say, what are the use cases that need to be solved? What other new use cases can we solve and how can we solve the existing use cases better? So I think we're going to continue walking down that. We've got a great community behind us. That's tremendously exciting to be a part of. Great. John, talk about some of the conversations. We were talking with Joe Madeleine, who's now a Google media analyst. There's always good, healthy debates around good progress. And so obviously storage is some block. You've got your block, you've got flash, you've got object storage, where the action is, we know that. What are some of the conversations that you guys are involved in? I mean, we had some commentary on Crouchhead earlier. Staff versus Swift versus, you know, all the things. What's your take on all that? It's a huge question. We could probably spend another hour talking about that. Here's the highlight. Reader's digest. So I think there's two big points on that. The first part is that within the world of looking at what applications are using, you've got a world that's moving from traditional filers and sands and things like that and embracing object storage. That's the big shift. It's not a particular object storage system. It's just, we need to make this shift so that we know that when we get an integrated product and just buy it off the shelf, that we know it speaks natively to object storage. And the advantage of that is the fact that, again, the object storage system, that software defined thing, is able to offload those hard problems so that applications can build out efficiency and their own value very well. Now, inside of this whole trend to object storage, you've got obviously a lot of different people providing different things. And my perspective on this is that we know that everybody has data. It's always growing. And you need to have ownership of everything that touches your data. And the only possible way you can do that is with open systems. And that's something that OpenStack gives you. It's not just the fact that you can look at the code, but the whole reason we're here is we have a community that is open that people can get involved with, the governance, the code, all of that. And so rather than trying to sling mud with other implementations of object storage, I think that people should be... Well, there's always religious wars, so to speak, technically, in communities, but the transition is real. I mean, object storage speaks to the megatrends that we're all living in. Devops, big data, integrated stacks, where applications are in control, so this needs some agility, really a big deal. That is just a fundamental approach. It's a mindset. So splitting the wars philosophies apart, where are we with object storage? Some of the key evolutionary innovation conversations happening. Is it just erasure codes, or is there anything else? I think that the promise that you get from object storage in that model is the fact that you can have your storage infrastructure actually responds and be configured very specifically to your use case. And that's something that you don't want something that's a one-size-fit-all solution that is very rigid and how you're deploying and how you're consuming the software. So what I think you're going to see is people who are able to take advantages of lack of hardware lock-in, but also open systems, and configure it to exactly their needs. That's the trend. I think the whole adaptive programming models of the years ago vision is happening right now, but it's happening in cloud in a different way. Super exciting. John Dickerson, thanks for coming on theCUBE. I want to give you the last word and share with the folks out there who want to know what's happening, who are here in Silicon Valley. What is the big thing happening here at OpenStack this year? What should they need to know about what's going on today? Thank you very much for having me. This has been a really great day, morning already here at the OpenStack SV thing. And I think that the big conversations continued this morning even within OpenStack today have to do a lot about figuring out how we move forward as a community and governance to scale it and accomplish everything that we can over the next four years. So we've learned a lot of lessons so far and figuring out what does it mean to be OpenStack and how does that all fit together is a crucial question to figuring out how we can both scale not only our individual projects, projects technically, but also looking at how do we scale the overall community so that we don't collapse under our own weight. And I think there's some really great ideas and things proposed right now. And I'm looking forward to continuing those discussions. And those are growing pains. Absolutely. It's a symptom of success, yes. Yeah, you're not getting smaller. It's a good problem, don't get mad. It's a hard problem, but yeah. Okay, John, take us in with SwiftStack here inside theCUBE, obviously a lot of great conversations open community, it's all happening in the sunlight and no black boxes anywhere. This is theCUBE, we'll be right back with our next guest after this short break. Thanks.