 So we're at SiliconANGLE.com, reference point for tech innovation. This is theCUBE, our flagship program, we go out to the events and extract the synth of the noise. I'm John Furrier, the founder of SiliconANGLE. I'm joined with my co-host. I'm Dave Vellante of Wikibon.org. Claude Barrera is here. He is the Chief Technical Strategist and Distinguished Engineer, IBM Storage Group. Where should we even start? Metadata. You have metadata, you have flash, you have metadata, we have software to design, we get open source, all the confluence of these megatrends going on. It feels like we've never seen this much before, maybe it just feels that way because it's new, but I don't know, what do you think? Well, it's definitely new, and let's sort of take that stack in order. We haven't seen this much innovation since Sands were new. And at that time, everybody rethought everything and we sort of went through how we were raised, built, and how are they connected, and how do people manage them, and what apps do you put on them. So we're going to go through all that again. Starting from the bottom, flash changes things. And just in time, because HDDs are really out of gas in terms of IOP performance. Chaplain yourself said, you know what, a lot of times you hear, oh, flash is going to give way to whatever, pick your persistent storage medium. And he said, I don't think so. I think it's going to be a hierarchy of semiconductor devices. Do you buy that, or do you think that one will win? It's possible, but we think one will win. And exactly which one in any particular year is kind of crystal ball stuff. We, the enterprise guys, don't make the decision. Decisions made by the consumer electronics people, the cell phone guys, the tablet guys, and our job in enterprise is to do technologies of aggregation. So we put the stuff together, make it not fail, make it maintainable, make it scale. Metadata was in the New York Times, the headline. Metadata, metadata, you know, this is now a mainstream term. People are now savvy to the fact that metadata is about information about information is out there, highlighted by the mainstream NSA story around Prism. And it's a big data story. So talk about metadata. Ambu said it's about understanding what you have, is metadata, metadata, and that the policy-based data management is going away. He believes, and it's going to be much more dynamic, much more intelligent. Can you comment from a technical perspective where we are in that journey of, the metadata has been around in storage for a while, but more importantly, as it becomes much more agile data, as it's more data processing going on at the edge of the network, whether it's Internet of Things, mobile devices, or just consumer data, policy-based data management, slow, static. Can you just talk about that and how you look at that? Well, so number one, it is good and exciting that people are waking up to the existence of metadata. It's an incredibly important technology and discipline. Metadata comes in lots of different sizes and shapes. And the answer to your question about how do we manage storage and what role does metadata play depends on what you're doing. So let's use a maybe less controversial example of how metadata might be used. If you're a large sporting organization that puts on football games every Sunday, you collect video of every game, every play from 20 camera angles. And so you've got mountains and mountains of video. But what's really useful is to have annotations that say in seven minutes and 20 seconds into the second quarter, so-and-so caught a pass thrown by so-and-so. And it went 25 yards. And here's the impact that it had on the game. And here's who the defender was and how he got beat. And that kind of stuff. Online gaming, too. Another first person went and took a shot. When they did, they traded currency. Gaming is a great example of Xbox announcing big things to me. So all these things where you create easily searchable information to layer on top of information that is not easily searchable. So all you had was video, finding the play where Guy X threw a pass to Guy Y is pretty hard. You need a human with eyes doing that. If you do annotation, and somebody has to do the annotation either automatically or a human, but once you've done the annotation, you now have data about data. And you can use that when it's time to go find the thing you want. But when you want to put the video clip on TV, you don't put up the metadata. You put up the video of the guy catching the pass. You're adding some twists. So let's start from the basics. Software Defined 101, what is it? Well, so you start with the idea that what you really want in terms of seeing storage is an abstraction. So we've talked about virtualization for a long time. You can think of that as step one in getting the underlying hardware. Abstracting the underlying hardware. So I see idealized behavior. I see homogeneous behavior across all my storage. I do snapshots in a single way. I can overprovision. I'm not constrained by the physicality of what the storage can do. OK, so that's progress. Now, the next step might be being able to dynamically create instances of that storage asset. So not only do I have this homogeneous abstraction, but if I decide right now, oh, I need a little more, or I need more bandwidth, I need to change the quality of service that it's providing, I can do that. So provision, capacity, and, say, performance, through an API call. That's right. Let's manage this through business policy. So rather than just being responsive to the demand, also have the notion of what really is the business asking for by rule. So yes, I want good response time, but this data can't move outside of my country. Or it has to be protected in some form of secure encrypted way when it moves outside of my data center. So it's not just responsive to immediate or perceived need, but here's the rules that you have to abide by within that response. And allowing full stacks, multi-vendor support, open standards, all those things, the tipping points of flash, and open collaborative. What's the one tech you're looking at that you're watching closely that everyone should be mindful for? Well, I'm not sure this is exactly what you had in mind. But the one tech that I'm watching because I think it's interesting, and I guess I think we don't really know all the places where it's going to be used, is Object Store. To me, the really interesting thing about Object Store is that they're simple. They're easy to program to, and they are likely to be the place that, say, mobile devices reach into to get whatever data the app on that mobile device knows it wants. Good for batch, a lot of batch. And real time. Good for immediate demand from, you know, you want something on your phone. There's a directory somewhere that knows where that object is. Go reach and get it. Pull it out. Excellent. Put get. Put get. OK. Simple, easy, and probably a lot of demand for it. Claude, thanks for coming on to theCube. We got wall-to-wall, non-stop, IBM events, Dave or Grave, all day long. We have a lot more to cover today. Day one of IBM Edge. And again, all day tomorrow. Stay here on SiliconANGLE.com. This is the Cube. We'll be right back with our next cast after this short break.