 and we're here with David Flynn, a many time guest on theCUBE. The CEO and founder, co-founder of Fusion I.O., David, welcome back. Thanks, David. Thanks, John. We've been on theCUBE many times. When you were a private company, now you're a public company, so we got to be careful in what you talk about because Fusion I.O. is officially public, so everything you can say will move the stock price, which is up today. You guys made some announcements, great to have you back. First question is, the world's changing, Oracle's now come into the market. Last year, the story's about cloud, and we talked about Oracle Open World. Last year, VMworld, we talked about cloud, but big data's a big part of their story today. At Java One, they're talking about MySQL and open source is gaining a lot of traction. Oracle's talking about performance. This is something that you're familiar with, performance, cloud, and big data. So what are you announcing today, and then we'll jump into what's going on with Fusion I.O.? So today we announced an extension to the storage primitives for use on Flash. It's been what, 30 years ago, that block read, block write came into existence, and that's sort of been the state of the art for the last 30 years. That worked great for mechanical hard drives and infrastructure based on that, but today you have more options because NAND Flash is fundamentally a new type of media and it has a different capability set. So the thing that we're introducing today is an atomic multi-block write that is native to the storage device. So, and this is being showcased in MySQL's N-O-D-B storage engine with the ability to get rid of the need to do a two step tap dance, I call it, to get an acid compliant transaction. You don't have to write it first to the log and then to the database in the correct place. You have to do that to avoid cutting it in half if you have a power outage or some other event. With an atomic write at the storage layer, the IO memory guarantees that it either writes all of it or none of it, so you don't have to sit there and double write everything. This increases the performance significantly because you're not having to write everything twice. It doubles the endurance of the Flash product, so it's a win all the way around. It also, most interestingly, simplifies the implementation within the application because now you don't have to implement the two face commit piece. So Larry, it's software, obviously we're talking about. It's software, yeah, this is an extension to our virtual storage layer. Fusion IO's architecture has been differentiated by the fact that our Flash is more like a memory device and the software stack is what teaches the operating system how to consume it using the traditional storage APIs. And what we're showing here is that it's not limited to just the traditional storage APIs, but you can have an additional set of much richer semantics. So Larry Ellison talked about performance all day long, Sunday night, and can't wait to hear what his keynote is on Wednesday because he tends to be more aggressive with his competitors, but he really talked about Oracle's vision around performance, parallelism. You've been leading the marketplace with performance, obviously with your product. What's the performance trends these days? Obviously what's going on in the market with cloud? Obviously with analytics, performance is real-time. What are you seeing as a CEO? You're out there talking to customers. You're in some of the biggest sites out there, powering performance. What are the big concerns? And what's the architecture frameworks for these new solutions? So you know, the value in data ages very quickly. It's not very valuable after very long. You know, you're in the content business, right? So the information economy is driven by our ability to take raw data and turn it into some sort of valuable knowledge that can be monetized and used quickly. And it's the quantity and rate at which you can do that processing that determines it. So, you know, clearly there's a huge demand for being able to do more faster in that area. So what FusionIO realized is that