 This is Dave Vellante at wikibon.org and this is theCUBE and we're back here at 590 Madison Avenue. And we've been talking today about how function, particularly storage function for the last 15 years has moved out of the server into the array, for good reason. To share, to replicate, to improve disaster recovery. But with Flash, we're starting to see function move back. As we've talked about wikibon and SiliconANGLE for quite some time, you really can't control fast data from slow storage. We're here with Colin Parris, who's the general manager of IBM's power systems business. We're going to talk about these and some other items, Colin, welcome to theCUBE. Glad to be here, Dave. Yeah, so we're hearing today about the big investments that IBM's making in Flash. And as I was saying in my introduction, we're starting to see servers and storage come together a little bit. So I want to talk about that. But first, give us an update on the power of business. I mean, it's a big business for IBM, driving a lot of transactions. What's going on in that business today? Ah, the power business is doing great. I mean, we're now at the point in which we've introduced the power seven plus, the newest generation of power, using some powerful new processors. We brought that out starting in October. We got the low end systems out in April and February. And so now we have the whole line of power in the marketplace. So it's all ready to go. Great against the competition like Sun and HP. And we've been doing fantastically well. Yeah, so a lot of databases running on power. Huge amount of databases. And Flash is the perfect complement to database. What are you seeing in terms of performance and just architectural changes that portend greater things for business value down the road? Well, I think two things matter to us here, right? So the first thing is the capability to do these online transaction processing. The OLTP, given the bandwidth that we see and the latency of Flash, we can do a lot of these huge transactions very, very quickly. So that's the first thing. The analytics that comes out of this as well is another astounding thing. What also matters to power quite a bit is the fact that we have a high level utilization. You know, normal Intel boxes run at 30%, 20% utilization. You can use power up to most clients, 60 to 90% utilization. That's great for us. Because with Flash right now, we can get the IO that allows a sustained utilization of all computing resources. So you buy less boxes and you buy power boxes. You don't want them running on these other machines that could only go up to 40%. So for us, it's the perfect combination. We get the much lower latency and we get to utilize these processes a lot more. So where are you seeing power uptake in the marketplace? Well, in terms of Flash, the two areas that come to mind right away, usually when you deal with things like any type of online trading, any type of transaction processing reservations, we see it right there. We see it in analytics, right? We've talked about people who are looking to find, for instance, intelligence agencies, once they get the data that quickly says where they use the analytics, they pull the data in, the data says, this is what's happening, now they want to run a large amount of computation on that data. We can run the computation very quickly and return a result. So it's not just getting the data off the database and doing the analysis, it's running the computation on that as equally as fast. So we're seeing a lot of people using the analytics for that reason as well. So if I think about just databases today, there's a lot of databases out there, they're relatively small and you try to minimize calls to the database because you don't want to go to the spinning disk if you don't have to and applications are somewhat limited by the amount of data that they can bring in and generally database systems are isolated from the rest of the systems in the organization. It sounds like given the comments you're making about analytics, that you see that changing, those two worlds coming together where transactional databases and analytic systems actually come together so that real-time decisions can be made. Are you seeing that today or is that more futuristic? No, we actually are seeing that today. The entire thing is to get the data as close as possible to the transactions. So you see it already in this in-memory view of the world. Now the interesting thing about power is that power, I think many people know it because of Watson, right? So we have the big, jeopardy machine. What most people know is that that was the largest in-memory database because what we had to do was we had to ensure that all of the data that was used in a jeopardy competition was on the Watson systems. It could not in any way connect to the internet. Largest in-memory database, you're pulling the data closer because you needed that response time. You needed that close speed. Flash begins to get you even closer, right? Normal disk, we're talking in terms of milliseconds. This is microseconds. Then you combine that capability with the in-memory capability. Now you have something truly unique. Everything is getting much faster, much more able to use the capabilities we have in terms of utilization. So we heard Steve Mills to say today that the discussion around in-memory has been around since memory has been around, but what hasn't been around is an extension of memory that's persistent. And we've sort of dreamed about this day for a number of years. How do you see that changing? Well, there's an old saying that we use at Wikibon. The best I.O. is no I.O. How do you see flash, that persistent storage medium, affecting memory as an extension? Do you see it going to the other side, what I would call the right side of the channel? Or do you see it as largely something that is still outside the channel and is really a storage medium? No, I quickly see it coming in. I think by and by, just the type of problems that we're going to address will require being there. If I go back 20 years ago, the discussion was all about the transaction, right? What am I doing to the transaction? Now the discussion is all about the question. How am I processing the question? So it was getting information was what you wanted out of a transaction making sure it was done. Now it's getting insights out of a situation. And the way to get those insights is to bring these things closer, to have all of your memory capabilities as close as possible. You see it even now in Hadoop systems in which you're trying to bring memory closer to the processing to do things with it, right? Now when I have flash capability and everything and in terms of the size is closer, there's some very different things I can do. And that is where I think transformation and industries begin when you can bring these things together. I'm delighted to start now because I think the people who are on it as early as possible get to transform industries. Yeah, now one of the things that IBM from my perspective has not done well is leveraged its expertise between specifically the storage group and the systems group. Now the organizational changes under Steve Mills and Rod, I think a lot of times companies, the organization is the biggest challenge. You've made those moves. Are you starting to work more closely with the storage group? Can we expect that you'll actually develop this end-to-end system? Definitely, because that's what you're seeing even now in this announcement, you notice I'm here, I'm running the Power Group. Ambudge is here, we're on storage. Pitchiano is here, we're on software. All of these pieces are coming together. And it's coming together for two reasons. One is clients demand it. So and we are clearly responsive that our client experience. The second reason it's happening is because they are now interdependencies that become very interesting. The SBC program that storage has, which gives you something known as easy tear. Now we have something that actually does easy optimization on Power as well. It begins to align the VM so that it can be as close to the cache as possible. So as you're aligning things to be close to the cache and you're aligning data to be close to memory and processors, wouldn't it make sense when those two things come together and that's exactly what's happening? You see some of the technical specialists we have here, the IBM fellows, they are working even closer together because now the fellow specialization is not about going deep, it's about going wide. So you'll notice here as well, you've got four or five fellows from storage, from Power, from software in terms of IM coming together to have these discussions. So we see this thing coming together on a fantastic rate. We've already done some things internally as well that make this happen. Also the other part that's interesting is in the go-to-market, we've begun to change things. Compensation schemes and ways that we actually allow the sellers to actually get paid for selling different things. All of that is beginning to come together as well. So I think you'll see some fairly profound changes. Yeah, the lines are blurring and for years companies like IBM have talked about how their heft and their integration, their global presence are an advantage and it has played out in certain cases but it really feels like it's starting to become a significant advantage. Talk about that a little bit. I think it clearly has been. If you look at some of the really cutting edge things we do, it has all the pieces in there. One thing I should say though, part of us playing out is also it playing out to the clients because when you go into some of the larger clients, they're organized in silos as well. It isn't coming together there. Now that we see in certain areas like analytics and cloud drawing people together in the organizations and sometimes in our clients, it makes it easier for us to come together because before if you went to the wrong person to do a sale, everything fell apart. So the industry itself is thinking about these things differently. If you notice, I mean in power, we also have something known as System Eye. So the legacy eye systems have always been about integration. We actually use that to leverage things like Pio. Pio has storage tied into it. So a lot of the products that are coming out now also are pulling us together. So it's both the industry doing it, the clients doing it, products doing it, our line men doing it. I think it's coming together. It will have to happen. So you're seeing the clients actually break down those organizations. That's happened over what the last 12, 24 months. Is that right or is it just starting to happen? I think in the last two years, I've seen it begin to take a different turn because issues like security that are showing up, issues like analytics and cloud force you to think together. You can't do it alone. You can't do a cloud in silos or analytics or security in silos. It just doesn't work. Okay, the lines as I say are blurring. As we've said many times, you can't run fast data from slow storage. And these worlds are coming together. Colin Parris, thanks very much for coming on theCUBE. It's a pleasure meeting you. Dave, it's a pleasure. Thank you very much. All right, keep it right there. We're right back with our next guest. This is Dave Vellante. This is theCUBE live from Madison Avenue.