Thank you and welcome. We appreciate you all coming out today. We've got a lot to cover, and we're starting a little bit earlier than normal. Normally, you probably know we do an afternoon, but as you probably have figured out by now, there was so much content, product, message, so much that came out. We actually had to spill some of this into the two weeks preceding this event, and we needed a little bit more time today to kind of walk through a lot of that with you. Let me just cover some of the housekeeping logistics items with you. First of all, we're going to do a little bit of lunch as part of the meeting. We'll break at one point. We'll ask you to go out, grab lunch.
It's a working lunch, folks, so please grab your food, come on back in, and we're just going to kind of keep it rolling. For those of you who haven't been before, the restrooms are right over here outside the room. If you could please use the doors on the back. We have some fire issues here on the side, so don't be using side doors, the back doors. We got power for everybody, so obviously be careful not to trip as you're kind of coming in and out of the rows. From a logistics standpoint, Wi-Fi, if you're having any problems, grab one of the guys towards the back of the room. We'll get that all sorted out. My big slide is always my favorite one, which kind of talks about Safe Harbor. Just some reminders here.
What I'm going to do here is I'm going to kind of cover some of these slides off. When the presenters are coming up and speaking with you, they're going to go by them pretty quick. The idea being here, they're going to reference back to here, this part of the meeting. Just as a reminder, we will be making forward-looking statements today and other information that kind of should be considered forward-looking. With that in mind, keep in mind that the statements that we're making, they are subject to risks and uncertainties, and that can cause actual results to differ from what we're talking about today. We're trying to present to you some important factors as we see them today relating to our business, but those statements can affect those forward-looking statements.
As a result, we caution you against placing undue reliance on these forward-looking statements because they reflect our opinions as of today. As a reminder, we're not obligating ourselves to update the slides that are being presented to you today, and we would encourage you to review our recent filings for a complete discussion of these risk factors and other things that may affect your decisions. Also, in addition to talking about forward-looking statements, we're going to be using GAAP and non-GAAP. Please keep in mind that we have a reconciliation on our website that you should look to to understand the important differences between what's presented on a GAAP basis and what's on a non-GAAP basis. Lastly, some of the product guys as they're talking about product technology want to make sure that you understand that the presentations being made today are for informational purposes only.
There's no forward commitments or promises being made. It's just that they're presenting to you. Thank you for your indulgence, and let me walk through those slides. As I mentioned earlier, over the course of the discussion, you're going to be seeing people reference those slides. They're not going to spend a lot of time on it. Let's get to the agenda. Welcome for being here. Safra is going to come on up and welcome you as well, especially as the host of the Financial Analyst Meeting. Afterwards, Mark's going to come up and speak about the strategy, talk about a lot of the stuff that came out this week, and there was a lot, no question about that.
Safra will come up and as our Chief Financial Officer, as well as one of our two Presidents, will speak with you on the financials and giving you an overview of Oracle's financial performance, which is quite strong, I might add. What we'll do is we'll give you 20 minutes to kind of do a little bit of housekeeping, run to the restroom, grab a lunch, come on back in the room, and then Thomas is going to come up on stage. Thomas is going to be here for a little bit because there's a lot of stuff that came out on the software side. He'll spend some time talking with you about that. He'll also be bringing up some folks from the specific units, whether it's database, middleware, or application, talking about some of the development activities in those spaces.
In that presentation, I'd like you to kind of keep in mind some things. I talked to a lot of you, as does Paul, Safra, Mark. We talked to you guys about a lot of stuff, and you bring up a lot of stuff. You know Oracle's not in the cloud. Okay, I think that kind of got answered. Oracle doesn't like big data. We are big data. You'll hear more about that. A lot of those things that you talk about, you're going to hear through Thomas, John, and all the other folks, Safra and Mark. You're going to hear from those folks that kind of lays out a lot of the things that you've been talking about. We're in it, and we're in it big. We'll do a Q&A with Thomas because I know you're going to have some follow-up questions for him on that.
We'll go to that. Then we'll jump to a break, another little bit of bio opportunity, 15 minutes or so. After that, we're going to come back into the back half of the day. We'll talk about support, Chuck Rozwat. We'll talk about license update and support on the software side, and also the hardware support opportunity that's quite good for us as well. We'll go to that. We'll go to Edward Screven. He'll come up and talk about infrastructure, things about Oracle VM or Linux, unbreakable Linux kernels. He'll talk to you about the operating system layer. He'll talk about that level. John Fowler is going to come on up and talk with you about much of the hardware. There's going to be a little bit of overlap between what John talks about and what Thomas talks about. Mark will hit this home point pretty clearly, as will Safra.
Hardware and software engineered to work together. There are natural overlaps. They're talking about it from each other's perspective. John will come there. What we're going to do is close out with those guys. We'll do a little bit of a Q&A. Any questions you have talking about hardware, support, or infrastructure, we'll do that. We'll bring Safra and Mark back up. I know you have a lot of questions for them. That's always good. We've got Larry is on the schedule here for end of day, but we'll talk about that in a moment. With that said, thank you very much for coming. Let me kick it over to Safra.
Hi, everyone. It's great to see you today. It is a bit of a somber day for us here at Oracle. What Ken was actually referencing, as many of you know, Larry and Steve are very close. We're very close friends. Our schedule is a bit fluid today, and I really won't be able to guarantee that Larry will end up being able to be here. I apologize in advance. The moment I know, one way or the other, I will let you know. Even when you're not surprised, as Larry wasn't, he'd seen Steve really just the day before or something like that, it's still a terrible shock. I apologize for our little bit of a crazy schedule today. With that, I want to really just take a moment to welcome all of you here. I am CFO again, and this time for keeps. I have embraced it.
I'm really hugging the role, as you can tell by the thousands of calls I've given you. Oh, no, I'm not allowed to do stuff like that. We have extended today. There's just a small quiz at the end. We are not going to tell you when the quiz comes, but there's like seven hours of material, and we've squished it as quickly as possible. We know a lot of you have travel plans, etc . We are going to go as quickly as possible. The truth is, we have so much to announce. We try to at least make it so that you don't have to go through the whole week of Oracle Open World to see what we've been up to.
I will have my opportunity to really chat with you at a little bit more length, and then, of course, really probably most importantly, to answer some of your questions after, as the day goes through. First, welcome. Thank you. I see a lot of familiar faces. I do want to make one bit of news right out of the box. We are not buying Yahoo. Just saving you the time on that. Other than that, let me introduce Mark Hurd. He was here the first time, actually, last year, just as he joined Oracle. We've had an absolutely clearly a wonderful year and a wonderful year together. I look forward to hearing his presentation. Mark Hurd.
Oh, thanks for that standing ovation. First pop quiz, how many of you consider yourself hostile? Good. Oh, really? Thank you. You want to stand up and declare yourself? Thank you. Yeah, thanks, Ken. Anybody here declares themselves as friendly? I see. So I see. Nobody wants to declare themselves at all. All right. Listen, thanks, Ken, for inviting me. I appreciate it. I'm not going to make any M&A proclamations. I only imagine where that might take me. Let's talk a little bit. We'll do a couple of things. One, tell you a little bit about where the company is, how the company's been doing. More importantly, tell you strategically what we're trying to get done and try to put a framework for when you hear a bunch of these things today.
You're going to hear a lot about technology, a lot about software and hardware, but try to give you a framework for how all this fits, what we're trying to get done in the marketplace with customers. Safra and I, at the end of the day, will come back and do Q&A with you about really any subject you want to hit, as long as it roughly has something to do with the IT industry. Listen, strategically, let me try to tell you how this all fits together. As you see the chart up here, when you look at this stack, we are trying to be best of breed at every layer of the stack. No matter whether it's the hardware layer, the middleware layer, the apps layer, we're trying to be best at everything we do. We line up from R&D through sales to execute that strategy.
High level of enterprise fit, fit into heterogeneous environments. Our customers can pick these capabilities à la carte. We want to gain share in every one of those categories. In addition to that, we vertically integrate these solutions in these things that manifest themselves in Exadata, Exalogic, and I'm going to talk more about that, to deliver an extreme level of both performance. Remember, performance and TCO are sort of one and the same. If you have a 100x performance gain and the solution costs $1 million, I can now buy that for $10,000 at the same level of performance. It's extreme change in sort of the dynamics of the industry, and it's incredibly disruptive. We do both. Everything isn't rotated. If you don't buy an engineered system, don't do business with Oracle. It's both. We deliver that through multiple delivery architectures.
I couldn't say earlier in the week, Oracle Public Cloud, which I can say today, public cloud. We'll help you build a private cloud. We'll do it on site. By the way, you can mix and match those solutions into whatever delivery vehicle is optimized for you. A region can be SaaS, public cloud. Another region can be on site. You can change your mind. Huge competitive advantage for the company. This portfolio is the most complete portfolio in the IT industry, end to end. There isn't a piece part in this portfolio that isn't best of breed or on its way over time with the roadmaps we've disclosed to best of breed. Next chart. Engineered systems. You may know we've sort of done these in chronological order. Exadata has sort of got a lot of pub. It's in many of the biggest accounts in the world.
We have a huge pipeline, fairly simple value proposition. It's engineered up through the database layer. You see examples, and you hear manifestations of 100x, 80x improvement in performance. That may sound like I've heard some of you on private calls tell me, you know, thanks for that. I do not believe you, which is OK. You should. This means a report that took 80 minutes now takes 1 minute. 40 minutes now takes 30 seconds. That sort of performance. It's pretty commonplace. To be honest with you, I'm a little worried we get a bit numb to it at Oracle. We hear 60x, 70x. In fact, if we ever hear 15x, somebody goes, what happened? What happened?
It really comes down to when we compress data, just even 5x, we get a 5x improvement in performance before we even start, let alone before we go to the massively parallel nature of the box. I'll let the engineering folks tell you more. Exalogic is on a better ramp than Exadata when it first started, engineered up through the middleware layer. You can see now we've released Exalytics, and we've released incremental additional engineered solutions. We don't call them necessarily Exa, the big data appliance. I'll focus maybe for a second on the Oracle Database Appliance. That is an x86. Yes, an x86 with the Oracle Database, Oracle Enterprise Linux that gives us a value-added intellectual property solution all from Oracle, differentiating the x86 market. There are 8 million x86 servers sold here. A couple million of them either are 2P, have Linux, or Oracle databases as a solution.
Relatively material market. We don't have interest in just selling an x86 with no other IP associated with it because none of that is our IP. With the stack I described, remember the average blade, if you will, on a $40,000 solution to a customer, the average blade is about $5,000. The rest of it is software. We now bring that portfolio that we can surround the technology with as an engineered system, all provisioned by us, all tested by us with one-button support, remote support on an x86. I've been doing this for a long time. Nobody has one-button support of an x86 solution with an integrated tested software stack. Next chart. We also this week released Fusion Applications, seven product categories, 100 products. They are GA and in the market. We're out with HCM. We're selling hard. We're on the trail with CRM today. We're on the attack.
I heard the CRM guys had a lot of comments yesterday. I did do one piece of financial analysis that I thought was interesting. When I looked at CRM's P&L, or if you will, Salesforce's, they make less money as a company than Safra makes in a year. Listen, as many of you know, I'm quite numerate. I actually studied this, and I think that this is correct. We're in the market, and that was not a joke. I'm very timid. Next chart. 200 customers, early adopters of Fusion, were on the attack with Fusion. Exa Fusion into the market in a big way in fiscal 2012. Next chart. In the Salesforce, we're doing three things at the same time. We're hiring more folks. We have more market opportunity than we have people. We're hiring people. Second, we're actually further specializing.
I mean, we believe, let me make clear, I believe the more specialized you are, the more productive you are, the more you sell. You have to know what you're selling. You have to know the problem you're solving. Specialization wins. We simultaneously are trying to deepen our relationships with our customers. These customers of ours are under a ton of pressure. Budgets, you guys may be investing in companies where CEOs are telling IT managers, whatever you need, you got. Most of them say, I got no money. The secular changes in the industry, how many of you guys are old enough to know what a VAX is? You ever heard of a DEC VAX? This is a big computer when a guy like me was just working. Cray Supercomputer, circa 1986. You heard of that? See this thing I got in my hand? DEC VAX.
Spec to spec, DEC VAX. 2 billion of them running around the planet, trying to access 65 zettabytes worth of data. Pressure on e-commerce, pressure on mobility, pressure on social networking, and all these strategies that our CEOs have to get into this new sector mainstream puts incredible pressure on IT budgets. We're a company that can come in and help them save money, as I mentioned earlier, and innovate with our application portfolio. Nobody else in the industry can do both at the same time. That is why we have to strategically engage with these customers at the strategy and business model level and the process level. Because companies' CEOs worry about strategy, strategy alignment to business model, aligned to process, aligned to apps, aligned to IT. That is why this third point about deepening these relationships is important to us.
We're doing all three at the same time: material growth in the Salesforce, more specialization, deeper relationships at a more strategic level. Next chart. I'll end on this. This is what you can focus us on or think of us working on in 2012. Organic growth. You saw it in Q1, 17%, sort of what Ken mentioned earlier. I think it's a pretty good number. Almost all organic, very little acquisition in those numbers. Scale out of Exa. We are going to scale out the things we got. We've just added more to it. Continue to focus on the Fusion Wrap and drive these industry. We did not talk about the industry groups as much today. They are a huge competitive advantage at a strategic level for us in accounts.
Retail, financial services, etc., have us in a conversation with CEOs about that very business model, business process, application change that I described earlier, and put us in a very advantageous competitive position. We want to continue to build those out. Listen, before I turn it back to Safra, I think we had a good year in 2010, not just financially, but also putting pieces together strategically. I told Larry a year ago, I just wasn't sure we could introduce more technology in a year than we did it this year a year ago. I think we just did it this week. You look at the way we filled out the portfolio, and I know when you look at our $32 billion worth of cash, there's always, what are you going to buy next? Where do you feel there's a hole in the portfolio?
I tell you, this portfolio right now is by far the best in the industry in trying to help enterprises get done what they got to get done. With what we've released through R&D, we've got a strong hand. We need to get, as I described, some of these actions I've described in the field. That is why we're hiring more people. We're focusing, training them more, at the same time investing in building out these relationships that I described. These industry groups are most helpful to us as well. I'm going to turn it back to Ken, to Safra. Safra is going to come back up. Thanks. We'll see you later.
All right. Late breaking news, Larry is coming. He will be here. That's why I had to take his call. Be easy on him. He is coming. He feels responsible to talk with all of you guys, and he really wants to get that opportunity. He will be here later today. Additionally, you're going to see a few of us popping in and out just a little bit. Don't be offended. We do have our own global managers meeting going on simultaneously. We're not like watching Housewives of New Jersey or anything, as much as I'd like to. OK, I don't think I have to do this again, right? Who remembers it? OK. My lawyers are seeing you. OK, great. We're in. As you know, we're using non-GAAP measures and GAAP measures. We try to so that you understand the business. You see it, frankly, the way we run it.
All right. Obviously, you know last year you've studied it. It was spectacular. The reason I bring it up now is because here is really the day where we start to put it in context. Obviously, new license was up big, really around the world. That is still the number I know, I hope you all look at. It is the way we add to our base. The software support number, as you know, is really not about support. It's about all of our users around the world. It is this massive base, nearly $15 billion last year, still growing. As you know, though many of you recommended against it, we still went into the hardware business. We were able to, instead of lose an absolute fortune, make an absolute fortune. Part of that was improving our gross margins dramatically. Our operating margins this past year were 44%.
I'll put that in a little bit of context for you. Earnings per share up 33%, not just driven by a massive, massive buyback like some of our competitors, but actually by growth. I know a number of you obsessed with operating cash flow or free cash flow, us too. It was up to $10.8 billion, up very substantially last year. We raised our dividend at the end of the year from $0.05 to -$0.06 a quarter, and it showed up as just $0.01 last year. It's obviously going to show up all year. Clearly a spectacular year. That includes, of course, a full year of Sun Microsystems. Interestingly, you'll see this is a period where we went through a global financial meltdown in the absolute middle. This is six years, includes the entire meltdown.
What you get is you really get to see that we have more than doubled as a company. We were already enormous in 2006. Yet, we're now really, really enormous now. New license revenue, even with a global financial meltdown showing up in the numbers, accelerated enormously in FY 2011. Very balanced, both in technology. I remember reading even 15 years ago that the database was dead. OK, not dead yet. Applications, again, passing our all-time high from before the meltdown, going past it this past year. As many of you saw, Q1, all cylinders firing. I would say especially in Europe, and I would say even especially in Europe, considering what we had in Q1. You see, very, very strong, very, very balanced around the world, both in the Americas, in EMEA, and in Asia-Pacific.
This is really a very important slide because this is, of course, what pays a lot of the bills for us, continues to grow at this absolutely shocking rate when the numbers are really so, so large. This is where you see our customers voting with their wallets. They are already customers, and they are buying more and more and more. They are depending on us to deliver more of their solutions. This is a sign of a very, very happy customer base. Obviously, Q1 was really a great quarter. It was particularly delightful for all of us because, as you know, Q1 for us is our seasonally smallest quarter, and it ends in August. I always say only engineers could think of a calendar fiscal year where Q1 ends in August as you chase around your customers to get them to sign software contracts.
For us, it was a very strong quarter. That is, even following a Q1 of the previous year where new license was up 25%, up again 17%, and EMEA obviously a standout, quite shocking, considering really all the turmoil that we just had in August. Software support up now to $4 billion in the quarter. Gross margins in hardware continuing to increase. Operating margins 360 basis points to 42%. For those of you, and some of you are actually in the audience who have covered Oracle for a very, very long time, you may remember that in Q1, Oracle didn't use to make any money. That was now the 1980s. Nonetheless, let's hope it doesn't take Salesforce.com another 20 years to actually make money. We increased our buyback. We increased our dividend.
I know some of you look at our hardware business, and clearly, you can see just a continuous improvement in the way we run the business. That is really a result of a lot of things. We are not in a commodity business ourselves. What we sell is our own intellectual property that is in the hardware. Customers are looking for a good combination of hardware and software, and they are willing to pay for it. As a result, we have left the business that Sun Microsystems embraced over the last few years of just sort of selling just to have those revenue numbers very high and selling the things that are of real value add. We have also, as I mentioned to you, reduced, and it's still, it's got to flow out the whole year, the resale of other people's products.
Obviously, those have very low gross margins in comparison. As those flow out, the hardware business will become even more profitable. Now, this one, this leading operating margins, I'm going to invite some of you later to explain to me why some software companies don't need to make money. We feel this incredible urge to send money back to our shareholders. Here we are, FY 2011, only 2 percentage points off our highest margins and actually the highest operating margins in the entire software industry. That includes an entire hardware company in us. Clearly, you can see right here that we are, and if you looked at Q1 we just had, there's absolutely, in my view, no doubt whatsoever that we will meet our previous highs and exceed them.
We are clearly well on our way to all-time high operating margins as a company, even though that includes an entire hardware company buried in us. Earnings per share growth, even through a global economic meltdown, I know it's getting kind of boring here, guys, but I'm going to keep going for the top to the right, a little higher, if you don't mind. Unless you tell me we don't need to make money, we're only going to spend like 1.5 times R&D on sales and marketing, not 5 times. You let me know if you want to switch that around. Until then, we're going to just keep focusing on sending money home to our shareholders. People ask us about what new stuff. Frankly, today is all about that. I mean, we spent $4.3 billion on R&D, OK? You're going to hear all about it today.
You're going to be exhausted by the time this is over. Frankly, we will probably have only touched the tip of the iceberg. You really cannot tell yet the impact of the software and the hardware and the software-hardware combination changes that we have made. Our goal, ultimately, as Mark said, our goal is actually to bring the absolute most value to our customers. As a result, they will reward us by buying our products. Frankly, during more difficult economic times, during a crisis or whatever is going on, it is actually driving folks to look at their existing standards and their enormous IT bills that they have, much of which is labor and other products, and saying, we've got to do something differently. That is what we're pushing.
Our goal actually is to change the industry, to make our products not 10% faster or better, not 20% faster or better, not even 100% faster or better. We are talking about 10 times better. We're talking sometimes about 100 times better. Our products, little by little, are turning into an IQ test, frankly. We have so, so many things to talk about that honestly, as it is, you probably can't even read the slides. Larry would never allow us to show a slide like this. It's got too much on it. It's simply a matter of so many things going on. Our hands are everywhere. Yet, we are growing in every one of these areas. These are very, very powerful things that you will only be able to see in your rearview mirror.
We will ask five years from now, wow, it's hard to believe that customers did it that old way, that vendors did not take full responsibility for providing them a complete service, a complete product. You'll see little by little. I know, see, I'm not a real PowerPoint person, as many of you know. I'm all about the spreadsheet. I'm right there with you guys. You don't have to believe me. Everything we've told you in the past has turned out exactly as it has. That is what's going to continue to happen here. We will continue to deliver you the best products. As a result, deliver to our customers the best products. They will reward us by voting with their pocketbook. These products, now we have stuff that we can actually show you.
We have a number of brand new products coming out that have come out, a number that you're going to start seeing more and more have an impact in the financial statements. We are very, very excited. The truth is that this is, in our view, the absolute best time at Oracle. I know I keep saying that, and it turns out to be so. The next year, I honestly cannot imagine a time when we have had as strong a portfolio and as strong a momentum. As you know, and as Mark mentioned, we are hiring aggressively, continuously. You all have the numbers. I feel silly even mentioning them to you. You see that our headcount is up. OK, why is that? Because, to be honest with you, we still have a lot of products that many of our customers don't even know about. It's almost overwhelming.
It's like they've walked into the Target store, and they don't know where to start. Our job is to make them walk down every single aisle and get everything they need. When they're in there, you know, they went in for batteries. They come out with boy socks and Doritos and a new movie. That's what it's like. At the checkout counter, they get some gum and a lint brush for some reason. At Oracle, it's pretty much the same thing. You come in for some database. You try on a little middleware, a little BI before you know it. While I'm here, I cannot tell you how many customers have said, you know, I was in. I was getting some BI. I was getting some middleware. I thought, well, I might as well get an Exalogic with that. That's what you're supposed to run it on, isn't it?
How is it that our tech numbers, which are such an enormous number on their own, continue to go up? What you don't realize is that the hardware is actually not the end of the train. It's actually just the beginning of the next train. As they start to see what they can do with the Exadata or Exalogic and now the new Exalytics and the big data appliance, all it does is it gets them deeper and deeper into understanding the power of the technology. Guess what they do? They buy more. They buy not only more hardware, but more software and more applications. This is an incredibly virtuous cycle we are on. It is only because customers are really recognizing the value. It's really one customer after another. We had our leader circle. We needed the big theater yesterday for our leader circle.
We had to kick some guy out for that room. I'll tell you, customer after another told me about their experiences. I'm sure you're talking to our customers, so you know that. That really translates ultimately into financial results. OK. That's pretty much the financial presentation. I'm going to take as many questions as you want, even with the press here, afterwards, if you can survive this. What I guess I'm asking is for you all to go get lunch, grab it, and come right back. You can go to the men's room. There must be three ladies in here starting in, and I'll be in there with you ladies. At what time? What? It's 11:15. You want to start at 12:00? It's 45 minutes. All right. Don't go real far, OK? All right. We'll see you in a few minutes, OK? You know what?
This is what we're going to do because we're coordinating global managers meetings simultaneously. I'm not going to have you sit here and eat lunch for 45 minutes. This is what we're going to do. I get to be in charge. This is so fabulous. Grab lunch, go to the men's room, and I'm going to take some Q&A, OK? If we can get Mark, he'll take it too right now. OK? Does that work for you all? Is that OK? Is that a plan? Thank you.
OK. Here's what we're going to do. I apologize. We had a little bit of a shuffle here. We kind of got through some of the stuff a little bit quicker than we planned. What we're going to go ahead and do now is we're going to go to Q&A with Safra Catz and Mark Hurd. Then what we'll do is it's going to take that till about 11:50 A.M. We're going to run, and then at 11:50 A.M., what we're going to do is have a quick break, allowing you the opportunity to kind of go out and get some food and basically sit down. We'll start with Thomas. We'll be starting at 12:00 noon, OK? We'll go ahead and do that. With that in mind, why don't I get Safra and Mark up on stage, and we'll get going?
Larry's got to move forward. OK, great. Sorry about all that. I thought I'd finished cooking, but the lunch wasn't actually ready. All I left them to do was add mustard and couldn't get it done. Sorry about that.
Come on in.
Come on in.
Are you busy running mice?
We ought to get the doors closed, probably.
Yeah, they're coming in. Sorry about that. See, in typical Oracle fashion, we're trying to save money, so we won't be feeding you. No, no, no. I promise lunch is coming. I apologize.
Why don't we close the doors because the acoustics are going to be?
No, that's all.
OK.
OK.
Go ahead. Yeah.
Hi. It's Adam from Morgan Stanley. I have, I guess, two little questions for Safra. One, you know, in previous analyst days, you've used the opportunity to talk about long-term earnings growth targets. Wanted to get your sense if you had any updated thoughts on that. Secondly, you are sitting on a lot more cash than usual. You bought back a lot of stock in the most recent quarter. How are you thinking about uses of capital going forward?
Sure. I mean, I still believe that 20% earnings per share growth is totally achievable for us. I actually believe that on an annual basis. We have just absolutely enormous opportunity in front of us. I think it is not even breaking through the water yet. You really can't see what's happening, but you're going to see it. We are extremely confident and comfortable with the business. I'm trying to sound very grown-up here, so I don't sound too overly effusive here. We are as strong as we have ever been. We are building momentum on all areas. We are putting our money where our mouth is by increasing distribution. We have the most massive product portfolio in the industry. I mean, yes, it's true. Microsoft also has an Exadata. They call it Xbox. We've got the only real one.
We've got the one that runs all those millions of Oracle workloads. There is absolutely no reason whatsoever. The truth is, other than a global financial meltdown, that compounded annual growth rate number would have been nearly 30%. In the middle of a meltdown, it's still over 20%. There is just no reason for us. Our targets are consistently in excess of that. Our potential, there's nothing between us and this goal line but our own feet. There is no one, no one standing in our way at all, as long as we don't trip and we don't mis-execute. So far, things have been going reasonably well. We're very upbeat. Adam, what was your second? Oh, all that money. OK. Larry says we can't just roll around in it in the boardroom. That was my idea. It's not actually ours. It's his and yours.
What we do is we send some of it back. We still don't have an absolutely complete footprint and things like that. We do opportunistic shopping. To the extent that last time there was a global financial meltdown, what do we do? We went shopping. We basically stole Sun Microsystems, I mean, bought Sun Microsystems out from under IBM. The reality is we look for opportunities, but we're not crazy with our money. We get it. It ain't our money. It's your money. It's Larry's money. We're very, very careful with it. We'll continue to, you know, we raise the dividend. We buy back more stock. We buy stuff. We're not a bank. We don't want to hold it. You do know, though, that some of that money is overseas. I'm definitely not going to just pay these unnecessary extra taxes on it to bring it back here.
We can still use it for our purposes there. We are very aware of it. It is a very large number right now. I'm trying to see if Larry will let us actually just roll around in it a little. That's it.
Why don't we go over to this side? Did you have one over here?
Thank you. Hi, it's Brad Zelnick with Macquarie. Mark, one quick question for you. In one of Larry's presentations earlier in the week, he talked about an updated target for 3,000 Exadata units this year.
Is that what he said?
I think I'd actually seen it up on a slide.
Yeah, I was running around other rooms, so I probably didn't see it.
OK. Regardless, obviously, a lot of confidence around the product. It's not surprising, speaking to your customers and partners this week, there's a lot of excitement driving tremendous value. Can you talk about what might have changed from what you told us at the end of Q4 after one quarter that gives you that much more confidence in anything in terms of the way you're going to take it?
Yeah, I wouldn't say anything's really changed. I think we had a lot of confidence in Q4. I think we've had confidence sort of all the way through. I think it's only been an issue for us of when the tipping points occur. Listen, remember one thing about Exa. We're selling a server, storage, software, and frankly, disintermediating a ton of services all with one solution. This is a disruptive, interrupting technology. You get all kinds of reactions, as you can imagine, when you interrupt the market the way we're interrupting those markets. Pipeline has never not been strong. It's strong. What happens is the more we sell, to be very blunt, the easier it becomes for us to sell more. When you look at our Q1, which was a very strong Q1 in both dimensions of what we sold and how our pipe continued to increase.
For us, there's never been a time we didn't have confidence. Directionally, we think we're going to have a really strong year. I think we're going to add to it. I think this week was important because last year we were talking about introducing the technology. We had a bit of a user base, but nothing like what we have now. Now you've got references. We have customers now that have tens and tens of these things with real business case studies about how these 70x numbers you hear are real. They can tell another customer. Now we're into architectural discussions with customers that are just, they're just more mature.
Now when you bring Exalogic into the dialogue and the fact we can move up to the middleware layer into the application layer, the introduction of Exalytics, the breadth of the Exa family, it's a strong, Safra's point, it is a strong portfolio we're bringing into the dialogue. I don't want to imply that there was some moment that we got increased. There's never been a moment we didn't have confidence and feel like we have a chance to really make something cool happen here.
Let me translate. Larry just can't keep a secret, OK? He's more private.
I'm trying to be a little bit, but yeah, I mean, we've never not felt like we could do it.
Yeah.
Hi, it's Brent Thill with UBS. Just in terms of your goals for 2012, there were two things I wanted to drill in. One was on sales capacity. You mentioned you're going to hire aggressively, but can you put that in the context of how much capacity you're going to add? Secondly, in the verticals, you've had a strong vertical footprint. Can you give us a sense of where you think you're going to be putting more energy in? Are there certain verticals that you're seeing that are really underrepresented at this point that you can make a lot of headway in 2012?
Yeah. In terms of sales capacity, think of 15%, 20% ballpark. I'm not going to give you an exact number, but it's roughly right. Lining that up with technical resource to support it because we need that technical presale support along with it. It's really a cross group. I want to make sure we're clear. We feel very good about the Exa family, but we're adding Infusion. We're going directly up at HCM. We're going directly at CRM. We're ramping up Infusion. We're ramping up in Exa. We're ramping up, frankly, in Database. We feel very good, to Safra's earlier point, almost when you go product line by product line. I believe what I told you, whatever it was an hour ago, that each part of our product line is in an investor breed position, sort of product by product.
We're adding really across the board emphasis in the areas that I described. The question about industry by industry, I mean, we're trying to strengthen all the industry verticals. We've had a pretty damn good run. There's not really, when we roll up the numbers, there's not an industry vertical that when I make those statements on the call about the industry verticals group faster than Oracle from a license fee perspective, it isn't dominated by one vertical group 4,000%, and the rest were flat. It's really sort of vertical by vertical. Some quarters, another one's ahead of another one. When you normalize it, we've had great strength in communications. We've had great strength in retail. We had great growth in utilities. Financial services is a market that we've been heavily investing into because it is the biggest vertical when you look at it as one singular vertical.
A place that bad news or good news, we don't have as strong a position in financial services as a percent of our revenue that maybe some other companies do. That we see as a tremendous opportunity. Where they're issued to take cost out, there is not a better solution in the world than Exadata, Exalogic to come help in some of those issues. It's a tremendous opportunity for us.
Yeah, Tim Sickler from Stifel. Exadata's had and Exalogic has had a lot of success. That's a new way to go to market. With cloud, you're bringing another way to market. Can you talk to us how you are going to manage the margins to that, sort of similar to how you managed with the Exa line and what maybe the margin goals there are?
For the margin for the cloud business?
Yep.
OK. I’ll tell you, the reality is that we actually think that the cloud business is extremely profitable, unlike some of the folks who are in that business who don't seem to make any money. For us, it's.
I have stats for you on how much money they made.
Yeah, no, I got it. Yeah. I don't need a stat for that. You told them already. Thank you. See, for us, remember, we do it all on our own technology. When you base it on our own systems, which are 10 - 100 times faster, you need a lot less energy. You need a lot less floor space. The way we run it and automate the entire system, we have a lot of experience with that. We've been doing a different type of cloud, what we call on demand, all the, you know, for many, many years ourselves. We have completely optimized that business. This is significantly simpler and more profitable. We actually, you know, we don't think it should be a business that makes no money. We actually think we can run it incredibly profitably.
We've been providing some of these different services and different versions for nothing because we feel like it really does not cost us very much to run. We do not expect the operating margins in that business to have any negative effect on my overall business, period. If I thought it would, I would never have told you, gee, we'll make it to 50%. I would have said, oh, now we're doing this, so it's done. No, no, no. We will pass 50%, OK? Maybe not this year, next year, whatever it is. We will absolutely get there. Each one of these little points, when your revenues are $30 billion, $40 billion, it's a lot of darn money. Yet somehow we are able to do it. There are massive economies of scale in our business.
We have an absolutely devoted and enormous customer base that continues to just vote with their pocketbook on whether we're providing them the services. The cloud business is what they are asking for in some areas. We're going to provide it for them. For those others who want to manage their own systems or who want to do what is potentially much more typical, which is a hybrid, some in, some out, that's how it's designed. There can be some in, some out. From an operating margin point of view, honestly, I am really, really confident we're going to make a lot of money, actually, a lot of money.
Remember the cost of goods sold model. We are going to have to find a partner to give us some applications, find a partner to give us some middleware, find a partner to give us a database.
We're going to do it.
I think so. I think the cost structure is going to be good too.
Yeah, that's right.
I think it's an important thing to get through people's head, though. When you're trying to build a cloud, you have to have the IP. You have to have the stack. Almost every one of the people you talk to has had to forge partnerships that have the transfer of wealth between them. In our case, we're bringing our stack, which gives us an incredible advantage. Add to it the distribution capability, and add to it the fact that the customer base is already so familiar with the technology. This is a nice business opportunity for us. We'll go back over here.
Thanks. It's Ricky Sandler from Eminence Capital. Can you guys tell us how you would think about a large acquisition at this point in the company's life cycle, say something bigger than $10 billion - $15 billion, maybe Safra from a financial standpoint, Mark from a strategic?
You have some in mind?
No, just you guys have been rumored to be buying everything under the sun, including your former company. I guess I just want to know how you'd think about it because when you talk about your product portfolio being as strong as it is, anything that large would clearly be some adjacencies or sort of change the complexion of the company a little bit. I just want to know how you think about it.
Let me tell you really the thought process. If it was something big, really big, we have to ask ourselves one thing. Remember, our absolute largest shareholder is not in this room right now, OK? I know a lot of you are big shareholders, but our absolute largest shareholder is not in this room, OK? He's got to look at his portfolio, which is us, and say, do I want to trade half or a quarter or whatever it is, depending on what it is? Do I want to trade out something? My whole thing is Oracle now. Do I want to trade some of it out for that, whatever that is? I just read another rumor this morning. Even my dad calls me and asks me if we're buying these things. He gets very worried, poor guy.
The reality is, I'm not sure that it's very clear that the shareholder that's not in this room, and many of you, I mean, if you want to buy that, go buy that. For us, there's got to be such a win-win combination that Mr. Ellison is ready to trade. If it's something really big, tens and tens of billion or whatever it is, he's got to decide that he's trading out. I'm not saying he's doing a stock, we're doing a stock deal, but the business that we have right now with so many positive drivers at every single level, does he want to trade out what we have really, and what he has spent 36, 34 years building? Does he want to trade that out for this other thing even partially? That's the way we think about it. That's what you ask. How do we think about it?
That's how we think about it. Would we rather have some of that than some of this? We like this, OK? We think this is compelling. We think we have opportunity with this, that there is nothing in our way right now. Trading it out for a whole bunch of other stuff that's not as unique, not really that interesting, frankly. It's got to pass an enormous hurdle. Don't believe everything you.
I'd also add that as good as, you know, you got to weigh it against as good as I think we are. There are distractions with stuff that comes like that. I want to make sure I emphasize how positive we feel about where we are. Something that would stop us from focusing on taking advantage of that wouldn't be something we'd sign up for. I'm not making a proclamation of any type, other than to say we need to get this hand played that we have and played very, very well. Doing something that fits strategically, it could be interesting. You also got to be able to run it operationally, and it's got to not take away from this opportunity we've got. Probably didn't give you everything you wanted, but that's OK.
OK.
OK.
Thank you.
Safra.
Cash?
Go ahead.
It's just got right here.
Thank you very much. Safra, [crosstalk] Merrill Lynch . Question for you, Safra. You said the 50% margin, that's truly commendable. What are the line items that you get leverage from as you go to Oracle? I have a question for Mark as well.
Sure. I mean, the line items obviously is first, just taking advantage of our own enormous scale, OK? Every time we buy something, every time we grow, we really just don't spend that much more money. So just A, everything, meaning the growth of the business. We have a backend infrastructure to manage a business, I believe, nearly twice as large, potentially. Secondly, we have a massive distribution capacity, which Mark and the team are making even bigger. Again, leverage economies of scale, not particular line item one way or the other. Maybe you could call that sales and marketing. You know, Oracle OpenWorld. I mean, whether we have twice as many customers or whatever, it still makes money. Yeah, thanks, Judy.
The rest of the business is obviously just operationally, the hardware business, continuing to run it well, growth in software license, again, a very, very profitable business, and ultimately just growing the installed base. This is such a large and important base of existing customers who pay upgrade fees to get the newest products from us. I mean, we leverage a $4.3 billion R&D budget. The marginal cost of goods sold of another piece of software is, as you know, very small. We have very, very high fixed costs. Growth is where it is. We literally see improvement, you know, really at every line item as we continue to scale. As the hardware starts to be sold, our newest products, we're very excited. We just announced, obviously, new microprocessors in new hardware, which will be available very soon. Orders are absolutely flooding in. We're really, we're thrilled.
We've already spent the money to build the R&D to build these. Now we get to sell them. Yes.
Mark, I know you guys sound pretty optimistic. This event is a wonderful opportunity for you to feel the pulse of the customer. Looking backwards, you guys have done really well. The macro headlines, is this something that customers are talking to you about? Is it even a factor in their IT budgets and how much they plan to spend with Oracle at this point in time? Just wondering what your take is, what's your feel for the pulse of those customers?
OK, I feel the same way I did on the conference call when that came up. We see a lot of momentum. I'm not here to tell you whether it's a broad-based macro momentum in the IT industry. I'm telling you as it relates to us, pipeline's pretty strong. I think remember, what I tried to say earlier was we come at it with two opportunities to help the customer. One is to innovate. The second is we can help the customer save money. When customers go through what they're going through right now, the opportunity for us to consolidate massive server farms, to consolidate big storage farms, and integrate them on Exadata, integrate their middleware layers on Exalogic, save them money, and deliver them incremental performance while we do it, disintermediate a ton of services in this kind of environment is also a strong play.
When you have that portfolio that we've been talking about, it allows you to go through whatever economic environment you've got because customers want to take advantage of both. We've got a strong hand. Our pipeline is strong. That's the way I'd answer it.
OK, Ken, can I ask you something? Should we be serving lunch right now?
In two minutes?
OK.
OK.
All right. Let's take one more question, lunch. We'll try to be back later for more questions if that works for you guys, OK?
Do you want to do one on this side?
Yeah, sure.
We will call it.
Yeah. Hey, John.
Hi, Safra. Thanks. This is John DiFucci from JPMorgan. Safra put up a slide, showed great growth on the top line, also even better growth on the bottom line with the leverage. It uses the word a lot. It's great to see. Cash sort of alluded to the buzz we're hearing, not just from Oracle, but from customers on the engineered system side. It sounds like things are going really well there. Safra's always been clear that we're going after profitable businesses. She's talking about it just about every question here tonight. At the same time, there's been, I think, some confusion from the investment community over sort of growth, right? A couple of quarters ago, the stock acted a little funky after a decent quarter, just because the hardware was a little bit weaker, even though the profitability in the hardware was really strong.
I hate to bring shaking your head no. This quarter too, hardware, it's doing sort of what you're saying it's going to do. We're going to drive profitability. We're going to drive these great products through here. Some of this commodity stuff is going to bleed off. My question, which I realize is less important now than it was two quarters ago, because I think you've done a good job explaining this to us and to the investment community, is when will the engineered products on the top line sort of more than offset sort of that slow bleed-off of those commodity products?
You know, I'll tell you, we're not in perfect agreement on when exactly because there was a lot of chaff in that business. They were big businesses. Their reselling storage was hundreds of millions of dollars, in fact, half a billion. You just kind of have to work our way through it. We have to do it twice because it wasn't only Hitachi, but then the whole LSI piece, which when it got sold to NetApp, we've been getting out of, right? That whole piece. There was still leftover X86. You've got to remember, we've owned this company for what, like a year and a half or something like that. We're really literally pushing down as much as possible the business that's not interesting for us. I think Q4 is a big compare. I think by Q4, I think you'll see it for sure. It could even happen earlier.
The truth is we've got Exadata, Exalogic, now SuperCluster just coming out. Remember, unlike with software, I got to build it. We only build it after you order it. Takes me a little while. Build
It now, Exalytics, and all the other engineered systems, too, by the way. We're in Q2. We have still a lot, you know, it's really a matter of compares. I think it will be, you will not be able to miss it at all next year. I mean, it's going to be a shocking situation. It's really hard to predict the future. It'll be much better when we meet up next year and we'll be kind of looking at the past. That'll be a little bit easier. We're.
I wish I could get you to, to people in some ways to quit thinking about it like that because it's just not the way we think about it. We think about growing the important parts of the product line. We think about getting the IP-based products growing. That's why we're transparent about how much growth you're seeing in exit, not just in product, but also in the pipe. That's where the real money is. The fact that we grow the attach, let me tell you, this attach is a far bigger deal than whether you had 1% growth or negative 1% on an aggregate hardware line, confused with a bunch of stuff that's got no IP associated with it. We're trying to be, the reason we were more transparent was to make sure you understood more how we're thinking about it.
We don't sit around in a room saying, we just got to get 1% growth on the hardware line. We talk about the growth of Exa. We talk about getting the M and T lines where we have true value add, differentiated IP, particularly now with T4. How do we then align that to attach in the long-term profitability of the aggregate ecosystem? The stuff where it was OEM, where we had virtually no IP, no attach, crummy margins, if it goes away faster, slower, it is what it is. That's why you hear this sort of, I don't know, you know, sort of message. Even with the x86, as clear as we've been, there are customers that still want to buy x86s from us. We actually raise the price. They keep buying from us. Sorry, I shouldn't have said that. They do.
Because they wanted us part of a solution. What we did with the Oracle Database Appliance, by integrating IP now, that becomes far more interesting to us because now our IP is part of the solution and we get all the attach. It's really not about the hardware issue. It's about the lines inside the hardware and what the trajectory is of each of those and the alignment of our aggregated ecosystem to it. I think what we're doing, we're going to be more and more clear and more and more transparent about where those lines are headed. When they fall into the aggregate growth line, they do. It's more important the trajectory of Exadata, of Exalogic, of the Oracle Database Appliance, etc. That's what we're thinking about.
Thanks.
Okay. That's it. Okay. Now they tell me lunch is actually out there. How about grabbing that and coming back in? We're doing Thomas Kurian and his team next. Okay. Thank you. All right. Late breaking news, Larry is coming. He will be here. That's why I had to take his call. It's easy on him.
We're going to go to Q&A with Safra Catz and Mark Hurd. With that in mind, why don't I get Safra and Mark up on stage and we'll get going?
Okay, great.
Go ahead. Yeah.
Hi, it's Adam from Morgan Stanley. I have, I guess, two little questions for Safra. One, in previous analyst days, you've used the opportunity to talk about long-term earnings growth targets. Wanted to get your sense if you had any updated thoughts on that. Secondly, you are sitting on a lot more cash than usual. You've bought back a lot of stock in the most recent quarter. How are you thinking about uses of capital going forward?
Sure. I mean, I still believe that 20% earnings per share growth is totally achievable for us. I actually believe that as, you know, on an annual basis. We have just absolutely enormous opportunity in front of us. I think it is not even breaking through the water yet. You really can't see what's happening, but you're going to see it. We are extremely confident and comfortable with the business. I'm trying to sound very grown up here, so I don't sound too, you know, overly effusive here. We are as strong as we've ever been. We are building momentum on all areas. We are putting our money where our mouth is by increasing distribution. We have the most massive product portfolio in the industry. I mean, yes, it's true. Microsoft also has an Exadata, they call it Xbox. We've got the only real one.
We've got the one that runs all those millions of Oracle workloads. There is absolutely no reason whatsoever. The truth is, other than a global financial meltdown, that compounded annual growth rate number would have been nearly 30%. In the middle of a meltdown, it's still over 20%. There is just no reason for us. Our targets are consistently in excess of that. Our potential, there's nothing between us and this goal line but our own feet. There is no one, no one standing in our way at all. As long as we don't trip and we don't mis-execute. So far things have been going reasonably well. We're very upbeat. What was your second? Oh, all that money. Larry says we can't just roll around in it in the boardroom. That was my idea. It could say it's not actually ours. It's his and yours.
What we do is we send some of it back. We still don't have an absolutely complete footprint and things like that. We do opportunistic shopping. To the extent that, last time there was a global financial meltdown, what do we do? We went shopping. We basically stole Sun Microsystems. I mean, bought Sun Microsystems out from under IBM. The reality is we look for opportunities, but we're not crazy with our money. We get it. It ain't our money. It's your money. It's Larry's money. We're very, very careful with it. We continue to raise the dividend, buy back more stock, and we buy stuff. We're not a bank. We don't want to hold it. You do know though that some of that money is overseas.
I'm definitely not going to just pay these unnecessary extra taxes on it to bring it back here, but we can still use it for our purposes there. We are very aware of it. It is a very large number right now. I'm trying to see if Larry will let us actually just roll around in it a little. That's it.
When we go over to this side, anybody, do you guys want over here?
Thank you. Hi, it's Brad Alan Zelnick with Macquarie. Mark, one quick question for you. In one of Larry's presentations earlier in the week, he talked about an updated target for 3,000 Exadata units this year. That's something.
Is that what he said?
I think I've actually seen it up on a slide.
I was running around to other rooms, so I probably didn't see it.
Okay. Regardless, obviously a lot of confidence around the product. It's not surprising speaking to your customers and partners this week. There's a lot of excitement driving tremendous value. Can you talk about what might have changed from what you told us at the end of Q4 after one quarter that gives you that much more confidence in anything in terms of the way you're going to take it?
Yeah, I wouldn't say anything's really changed. I think we had a lot of confidence in Q4. I think we've had confidence sort of all the way through. It's only been an issue for us of when the tipping points occur. Remember one thing about Exadata. We're selling a server, storage, software, and frankly disintermediating a ton of services all with one solution. That's a disruptive, interrupting technology. You get all kinds of reactions, as you can imagine, when you interrupt the market the way we're interrupting those markets. The pipeline has never not been strong. It's strong. What happens is the more we sell, to be very blunt, the easier it becomes for us to sell more.
When you look at our Q1, which was a very strong Q1 in both dimensions of what we sold and how our pipe continued to increase, for us, there's never been a time we didn't have confidence. Directionally, we think we're going to have a really strong year. I think we're going to add to it. I think this week was important because last year we were talking about introducing the technology. We had a bit of a user base, but nothing like what we have now. Now you got references. We have customers now that have tens and tens of these things with real business case studies about how these 70x numbers you hear are real. They can tell another customer. Now we're into architectural discussions with customers that are just more mature.
Now when you bring Exalogic into the dialogue and the fact that we can move up to the middleware layer into the application layer, the introduction of Exalytics, the breadth of the Exadata family, it's a strong, to Safra's point, this is a strong portfolio we're bringing into the dialogue. I don't want to imply that there was some moment that we got increased. There's never been a moment we didn't have confidence and feel like we have a chance to really make something cool happen here.
Let me translate. Larry just can't keep the secret. He's more credible.
I'm trying to be a little bit, but yeah, I mean, we've never not felt like we can do it.
Hi, it's Brent Thill with UBS. In terms of your goals for 2012, there were two things I wanted to drill in. One was on sales capacity. You mentioned you're going to hire aggressively, but can you put that into context of how much capacity you're going to add? Secondly, in the verticals, you've had a strong vertical footprint, but can you give us a sense of where you think you're going to be putting more energy in? Are there certain verticals that you're seeing that are really underrepresented at this point that you can make a lot of headway in 2012?
Yeah. In terms of sales capacity, think of 15%, 20% ballpark. I'm not going to give you an exact number, but it's roughly right. Lining that up with technical resource to support it because we need that technical pre-sale support along with it. It's really a cross group. I want to make sure we're clear. We feel very good about the Exa family, but we're adding Infusion. We're going directly up at HCM. We're going directly at CRM. We're ramping up Infusion. We're ramping up in Exa. We're ramping up, frankly, in Database. We feel very good, to Safra's earlier point, almost when you go product line by product line, I believe what I told you, you know, whatever it was an hour ago, that each part of our product line is in an investor-breed position, sort of product by product.
We're adding really across the board emphasis in the areas that I described. Your question about industry by industry, I mean, we're trying to strengthen all the industry verticals. We've had a pretty damn good run. There's not really, when we roll up the numbers, there's not an industry vertical that, when I make those statements on the call about the industry verticals grew faster than Oracle from a license fee perspective, it isn't dominated by one vertical grew 4,000% and the rest were flat. It's really sort of vertical by vertical. Now, some quarters, another one's ahead of another one, but when you normalize it, we've had great strength in communications. We've had great strength in retail. We had great growth in utilities. Financial services is a market that we've been heavily investing into because it is the biggest vertical when you look at it as one singular vertical.
A place that, you know, bad news or good news, we don't have as strong a position in financial services as a percent of our revenue that maybe some other companies do. That we see as a tremendous opportunity. Where they're issued to take cost out, there is not a better solution in the world than Exadata, Exalogic to come help in some of those issues. It's a tremendous opportunity for us.
Yeah, Tim Sickler from Stifel . Exadata has had, and Exalogic has had a lot of success, and that's a new way to go to market. With cloud, you're bringing another way to market. Can you talk to us how you are going to manage the margins to that, sort of similar to how you managed with the Exa line? What maybe the margin goals there are?
For the margin, for the cloud business?
Yep.
Okay. I'll tell you, the reality is that we actually think that the cloud business is extremely profitable. Unlike some of the folks who are in that business who don't seem to make any money. For us, it's.
I have stats for you on how much money they make.
Yes, no, I got it. Yeah. I don't need a stat for that. You told them already. Yeah, thank you. You see, for us, remember, we do it all on our own technology. When you base it on our own systems, which are 10 to 100 times faster, you need a lot less energy, you need a lot less floor space. The way we run it and automate the entire system, we have a lot of experience with that. We've been doing a different type of cloud, what we call on demand, for many, many years ourselves. We have completely optimized that business. This is significantly simpler and more profitable. We actually, you know, we don't think it should be a business that makes no money. We actually think we can run it incredibly profitably.
We've been providing some of these different services in different versions for nothing because we feel like it really does not cost us very much to run. We do not expect the operating margins in that business to have any negative effect on my overall business. Period. If I thought it would, I would never have told you, gee, we'll make it to 50%. I would have said, oh, well, now we're doing this, so it's going to, no, no, no. We will pass 50%. Maybe not this year, next year, whatever it is, but we will absolutely get there. When it's, you know, each one of these little points, when your revenues are $30 billion, $40 billion, it's a lot of darn money. Yet somehow we are able to do it. There are massive economies of scale in our business.
We have an absolutely devoted and enormous customer base that continues to just vote with their pocketbook on whether we're providing them the services. The cloud business is what they are asking for in some areas, and we're going to provide it for them. For those others who want to manage their own systems or who want to do what is potentially much more typical, which is a hybrid, some in, some out. That's how it's designed, so there can be some in, some out. From an operating margin point of view, honestly, I am really, really confident we're going to make a lot of money, actually, a lot of money.
You know, remember the cost of goods sold model. We're going to have to find a partner to give us some applications, find a partner to give us some middleware, find a partner to give us a database.
We're doing great with us.
I think so. I think the cost structure is going to be good too.
Yeah.
When we negotiate on it, it goes well. I think it's an important thing to get through people's head though, when you're trying to build a cloud, you have to have the IP. You have to have the stack. Almost every one of the people you talk to has had to forge partnerships that have the transfer of wealth between them. In our case, we're bringing our stack, which gives us an incredible advantage. Add to it the distribution capability, and add to it the fact that the customer base is already so familiar with the technology. This is a nice business opportunity for us. We'll go back over here.
Thanks. It's Ricky Sandler from Eminence Capital. Can you guys tell us how you would think about a large acquisition at this point in the company's lifecycle? Say something bigger than $10 - $15 billion, maybe Safra from a financial standpoint, Mark from a strategic and.
You have something in mind?
I just, you know, you guys have been rumored to be buying everything under the sun, including your former company. I guess I just want to know how you'd think about it, because when you talked about your product portfolio being as strong as it is, anything that large would clearly be some adjacencies or sort of change the complexion of the company a little bit. I just want to know how you think about it.
Let me tell you really the thought process. If it was something big, you know, really big, we have to ask ourselves one thing. Remember, our absolute largest shareholder is not in this room right now. I know a lot of you are big shareholders, but our absolute largest shareholder is not in this room. He's got to look at his portfolio, which is us, and say, do I want to trade half or a quarter or whatever it is, depending on what it is? Do I want to trade out something? You know, I own my whole thing is Oracle now. Do I want to trade some of it out for that, whatever that is? I just read another rumor this morning. Even my dad calls me and asks me if we're buying these things. He's very worried, poor guy.
The reality is, I'm not sure that it's very clear that the shareholder that's not in this room, and many of you, I mean, if you want to buy that, go buy that. For us, there's got to be such a win-win combination that Mr. Ellison is ready to trade. If it's something really big, tens and tens of billion or whatever it is, he's got to decide that he's trading out. I'm not saying he's doing a stock, we're doing a stock deal, but the business that we have right now with so many positive drivers at every single level, does he want to trade out what we have really, and what he has spent 36, 34 years building? Does he want to trade that out for this other thing, even partially? That's the way we think about it. That's what you ask. How do we think about it?
That's how we think about it. Would we rather have some of that than some of this? We like this. We think this is compelling. We think we have opportunity with this, that there is nothing in our way right now. Trading it out for a whole bunch of other stuff that's not as unique, not really that interesting, frankly. It's got to pass an enormous hurdle. Don't believe everything you.
I'd also add that as good as, you know, you got to weigh it against as good as I think we are. There's distractions with stuff that comes like that. I want to make sure I emphasize how positive we feel about where we are. Something that would stop us from focusing on taking advantage of that wouldn't be something we'd sign up for. I'm not making a proclamation of any type, other than saying we need to get this hand played that we have and played very, very well. Doing something that fits strategically, it could be interesting, but you also got to be able to run it operationally, and it's got to not take away from this opportunity we've got. Probably didn't give you everything you wanted, but that's.
Okay.
Okay.
Okay.
Thank you.
Cash, go ahead. It's this guy right here. Go ahead.
Thank you very much, Safra. [crosstalk] . Question for you, Safra. You said the 50% margin, that's truly commendable. What are the line items that you get leverage from as you go towards it? I have a question for Mark as well.
Sure. I mean, the line items obviously is first just taking advantage of our own enormous scale. Every time we buy something, every time we grow, we really just don't spend that much more money. A, everything, meaning the growth of the business. We have a backend infrastructure to manage a business, I believe, nearly twice as large, potentially. Secondly, we have a massive distribution capacity, which Mark and the team are making even bigger. Again, leverage economies of scale, not particular line item one way or the other. Maybe you could call that sales and marketing. Oracle OpenWorld, I mean, whether we have twice as many customers or whatever, it still makes money. Yeah, thanks, Judy.
The rest of the business is obviously just operationally, the hardware business continuing to run it well, growth in software license, again, a very, very profitable business, and ultimately just growing the installed base. This is such a large and important base of existing customers who pay upgrade fees to get the newest products from us. I mean, we leverage a $4.3 billion R&D budget, and the marginal cost of goods sold of another piece of software is, as you know, very small. We have very, very high fixed costs, and just growth is where it is. We literally see improvement really at every line item as we continue to scale. As the hardware starts to be sold, our newest products, we're very excited. We just announced, obviously, new microprocessors in new hardware, which will be available very soon. Orders are absolutely flooding in. We're really, we're thrilled.
We've already spent the money to build the R&D, to build these. Now we get to sell them. Yes.
Mark, I know you guys sound pretty optimistic, and this event is a wonderful opportunity for you to feel the pulse of the customer. Looking backwards, you guys have done really well with the macro headlines. Is this something that customers are talking to you about? Is it even a factor in their IT budgets and how much they plan to spend with Oracle at this point in time? Just wondering what your take is, what's your feel for the pulse of that?
You know, Cash, I feel the same way I did on the conference call when that came up. We see a lot of momentum. I'm not here to tell you whether it's a broad-based macro momentum in the IT industry. I'm telling you as it relates to us, pipeline's pretty strong. I think, remember, what I tried to say earlier was we come at it with two opportunities to help the customer. One is to innovate. The second is we can help the customer save money. When customers go through what they're going through right now, the opportunity for us to consolidate massive server farms, to consolidate big storage farms and integrate them on Exadata, integrate their middleware layers on Exalogic, save them money, and deliver them incremental performance while we do it. Disintermediate a ton of services in this kind of environment is also a strong flight.
When you have that portfolio that we've been talking about, it allows you to go through whatever economic environment you've got because customers want to take advantage of both. We've got a strong hand. Our pipeline is strong, and that's the way I'd answer it.
Okay, Ken, can I ask you something? Should we be serving lunch right now?
In two minutes.
Okay. All right. Let's take one more question, lunch, and we'll try to be back later for more questions if that works for you guys. Okay.
You want to do one on this side?
Yeah, sure.
We will call it.
Yeah, hey John.
Hi, Safra. Thanks. It's John DiFucci from JP Morgan. Safra put up a slide, showed great growth in the top line, also even better growth in the bottom line with the leverage. She uses the word a lot. It's great to see. We, by cash, sort of alluded to the buzz we're hearing, not just from Oracle, but from customers on the engineered system side. It sounds like things are going really well there. Safra's always been clear that we're going after profitable businesses. She's talking about it just about every question here tonight. At the same time, there's been some, I think, some confusion from the investment community over sort of growth, right? A couple of quarters ago, the stock acted a little funky after a decent quarter, just because the hardware was a little bit weaker, even though the profitability in the hardware was really strong.
I hate to bring shaking your head no, but this quarter too, you know, hardware, it's doing sort of what you're saying it's going to do. We're going to drive profitability. We're going to drive these great products through here. Some of this commodity stuff is going to bleed off. My question, which I realize is less important now than it was two quarters ago, because I think you've done a good job explaining this to us and to the investment community, is when will the engineered products on the top line sort of more than offset sort of that slow bleed off of those commodity products?
You know, I'll tell you, we're not in perfect agreement on when exactly because there was a lot of chaff in that business, but they were big businesses. Their storage, reselling storage, was hundreds of millions of dollars. In fact, it's $500 million. You just kind of have to work our way through it, and we have to do it twice because it wasn't only Hitachi, but then the whole LSI piece, which, when it got sold to NetApp, we've been getting out of, right? That whole piece, then, there was still leftover x86. I mean, you got to remember, we've owned this company for what, like a year and a half or something like that. We're really literally pushing down as much as possible the business that's not interesting for us.
I think Q4 is a big compare, but I think by Q4, I think you'll see it for sure. It could even happen earlier. The truth is we've got Exadata, Exalogic, now SuperCluster just coming out. Remember, unlike with software, I got to build it. We only build it after you order it. It takes me a little while to build it. Now Exalytics and all the other T systems too, by the way. We're in Q2. We got still a lot, it's really a matter of compares. I think it will be, you will not be able to miss it at all next year. It's going to be a shocking situation. What is it? It's really hard to predict the future. It'll be much better when we meet up next year and we'll be kind of looking at the past. That'll be a little bit easier. You know, we're.
I wish I could get you to people in some ways to quit thinking about it like that, because it's just not the way we think about it. We think about growing the important parts of the product line. We think about getting the IP-based products growing. That's why we're transparent about how much growth you're seeing in Exa, not just in product, but also in the pipe. That's where the real money is. The fact that we grow the attach, let me tell you, this attach is a far bigger deal than whether you had 1% growth or negative 1% on an aggregate hardware line confused with a bunch of stuff that's got no IP associated with it. We're trying to be, the reason we were more transparent was to make sure you understood more how we're thinking about it.
We don't sit around in a room saying, we just got to get 1% growth on the hardware line. We talk about the growth of Exa. We talk about getting the M and T lines where we have true value add, differentiated IP, particularly now with T4. How do we then align that to attach in the long-term profitability of the aggregate ecosystem? The stuff where it was OEM, where we had virtually no IP, no attach, crummy margins, if it goes away faster, slower, it is what it is. That's why you hear this sort of, I don't know, you know, sort of message. Even with the X86, as clear as we've been, there are customers that still want to buy X86s from us. We actually raised the price. They keep buying from us. I'm sorry, I shouldn't have said that. They do.
Because they wanted us part of a solution. What we did with the Oracle Database Appliance, by integrating IP now, that becomes far more interesting to us because now our IP is part of the solution and we get all the attach. It's really not about the hardware issue. It's about the lines inside the hardware and what the trajectory is of each of those and the alignment of our aggregated ecosystem to it. I think what we're telling you, we're going to be more and more clear and more and more transparent about where those lines are headed. When they fall into the aggregate growth line, they do. It's more important the trajectory of Exadata, of Exalogic, of the Oracle Database Appliance, etc. That's what we're thinking about.
Thanks.
Okay. That's it. Now they tell me lunch is actually out there. So have a grab in that and come back in. We're doing Thomas Kurian and his team next. Okay. Thank you.
Welcome back, everybody. Appreciate the quick turn on the lunch. I know that we had a little bit of a quick on the fly, some adjustments there. Thank you for your patience and consideration on that. Glad we were able to get the Q&A with Mark and Safra in. As Safra pointed out, we'll work to see if we'll stay flexible with this. Obviously, with the events happening yesterday, things are a little bit more fluid than they might normally otherwise be. What we're going to go ahead and do now is we're going to shift the conversation to start to get into more of the specific technologies that we're working with. We'll go now into software, but clearly there's elements into integrated systems and engineered systems, as you'll hear from Thomas. Let me bring to the stage now Executive Vice President Thomas Kurian.
Thank you very much. What I'm going to try and do is to give you context of all of the different product businesses we're in and what is our model for differentiation and competitive strategy in those businesses. Through the course of it, I'll cover a lot of the announcements that we made at Open World and give you context for the rationale behind those announcements as well. I'm going to start with just the safe harbor statement. I'm not going to read it, but getting to our product strategy on the software side of the business, there's four basic elements to our product strategy. To deliver a complete, open, integrated—by integrated we mean we engineer the software—and best-in-class suite of database, middleware, and business applications. Second piece, to deliver engineered systems, combining our software and hardware together to deliver a breakthrough organic innovation.
Third is to deliver Oracle's technology and business applications, coupled with content and partner solutions through Oracle Public Cloud, which we announced yesterday. Lastly, to leverage our scale and install base and product breadth to upsell and cross-sell a broad range of integrated solutions to customers. We are in a variety of different segments of the business: storage, servers, operating systems, virtualization, and then above that, database, middleware, and business applications. Our goal is to offer customers a complete suite in every slice of this vertical stack. A complete suite of database technology, a complete suite of middleware, a complete suite of business applications, for example, based on open standards. When we say open, it means two important elements.
First of all, we've always had a history as Oracle of building based on open standards and to develop our products based on open standards, and customers like that because it gives them choice. The second is that our software is explicitly tested to work in a heterogeneous environment, combining Oracle and non-Oracle systems and solutions. The value to that is it makes it easy for customers to adopt a piece of our stack without having to go through a complete rip and replace in order to pick up a piece of our stack. Integrated is something we've done for a long period of time, which is we've engineered different pieces of our product line to work together particularly well. Whether it's different portions of our database offering, where we've engineered pieces to work together, different pieces of our middleware, and the ability to offer a single integrated middleware stack.
Also, with our business applications, as you hear about the introduction of Fusion Applications, one important thing we've done is to allow customers to coexist pieces of Fusion Applications with their existing footprint. That allows us to both bring our technology to a broadest possible install base of customers. For example, it allows a large number of customers to look at Fusion and decide that they want to pick up pieces of it. It also gives them value because the more pieces of our portfolio that you use, there's a lot less cost in both development as well as maintenance and operations. Finally, best in class. To leave it in a nutshell, no other vendor in the enterprise software space is a leader in as many technology and applications market segments as we are.
Not only that, no software vendor in the enterprise software space has added in the last year as many leadership segments as we have. We are not offering customers the choice of having either best of breed or integrated. It is best in class in every part of our product segment and engineered together. Okay, what I'm going to do now is to focus on the three primary business segments: database, middleware, and business applications, and go through the product strategy needs segment in more detail. Starting with the Oracle Database.
Our strategy with the Oracle Database is to deliver market-leading transaction processing, data warehousing, and big data management solutions with really breakthrough advances in performance, scalability, high availability, and security, coupled with unique software innovations optimized for new types of hardware architectures so that customers can get really significant improvements in both performance and reliability, but getting it at much lower cost. We talk about enabling customers to get better performance and reliability and doing that by ensuring that they spend less money to get that. Now there are seven different areas of the product line that we work in. Storage management, which is how we're optimizing the way databases store data and retrieve them from disk. Grid or cluster computing, which is the ability to support both scale-up and scale-out with very high performance and easy-to-manage systems. Then transaction processing and data warehousing.
Maximum availability, which is giving you the highest levels of reliability on open systems. The ability to manage both what in the market is called big data, which typically means large volumes of structured and semi-structured data in a single system. Additionally, we offer solutions to manage these environments with our database management technology and to protect the data when it's inside the database, as well as in flight into the database with our security options. Let's go through a few of these areas. Before I do that, the market share numbers I wanted to point out is that these are calendar year 2009 numbers, which showed that Oracle had 48% market share in overall database. In data warehousing, it had 39.6% market share. The reason I point that out is you are free to look at Gartner and IDC's latest numbers for 2010.
There are two important points I wanted to point out there. Not only did the Oracle database grow faster than the overall market for databases, not only did it grow faster in every segment, transaction processing and data warehousing, but the growth in the Oracle database market share was greater than all the other competitors combined. We are not only the market leader in this segment, we are the fastest growing, and we are growing market share faster than all the other players in the market combined. Here are the numbers, and you are free to, all these reports are publicly available, and you can look at them. Why is it that so many customers choose our database product?
It is because of the fundamental innovations we have made in the database, year after year, release after release, and the fact that we are looking several years ahead of the competition in identifying areas where customers would want us to offer some fundamental innovation and then architecting it into the product. Let me walk through some of the areas of investments that we focus in. Storage optimization. We started this effort because as storage, as the volumes of data in an enterprise has grown, the cost of storage, frankly, is increasingly a bigger part of the overall IT footprint. In many customers, it's actually even larger than the cost of servers and the investment that they make in servers. Okay, the work that we did spans a couple of different areas. Automatic storage management.
What this is, is basically the Oracle Database manages how data is written to disk and automatically balances the data across a cluster of NAS or SAN storage to ensure that both we make the application that's accessing the database optimal performance. We eliminate the need for database administrators to have to do this manual task. We save cost on storage by being most efficient in the way we're deploying storage blocks on disk. Hybrid columnar and advanced compression allows customers to get between 4 - 12 times compression of the data on disk and to do so without having impact on application performance. In contrast with other forms of storage compression from other vendors, what we do with advanced compression and hybrid columnar compression is allow you to compress data at the storage level while giving you excellent performance from a query and transaction performance in the database.
Cluster computing is really about reducing server costs. This is allowing you to put a pool of servers together. By putting the pool of servers together, you get the ability to get better performance the more computers you add into the cluster, at the same time getting better reliability. Other cluster solutions say you can either get better performance into the system by adding computer nodes into the cluster, or you get better reliability, but you can't get both. We give you that. It has been an area for a number of years where we've made innovations in there. The reason I pointed these two out is automatic storage management and cluster computing are the heart of what we do with Exadata.
In order for another vendor to offer a solution like Exadata, they would have to go back in time five years ago and build all of these technology solutions because Exadata does not run a product like Exadata could not be possible without automatic storage management, without columnar compression, without cluster computing, etc. More recently, we've added newer solutions. We've added capabilities to handle unstructured data. We support very large volumes of both structured and unstructured data in the system using capabilities like partitioning. At Oracle Open World, we announced a new product called the Oracle NoSQL Database. The NoSQL Database is a database that we offer. It's built from the Berkeley, it's built on top of Berkeley DB.
It gives you the ability for customers running unstructured data like log files and things coming off of web servers and other things to use a key value-based persistence model and get very efficient query performance from the system. We are broadening our footprint continually in the database, managing new types of data, and giving customers the same performance scalability advantages that we've traditionally offered. In memory, what we're really doing with in-memory technology is to bring it into three or four different parts of our product footprint. First of all, with the Oracle Database, we provide a number of capabilities to handle in-memory both OLTP and data warehousing. Andy Mendelsohn is going to talk about what we've done with Exadata. We also have TimesTen, which is used for in-memory OLTP.
At Oracle OpenWorld, we introduced a new version of TimesTen, specifically as part of our business intelligence portfolio to support in-memory analytics with a number of new features in that that I'll talk about when we get to Oracle Exalytics. We also have a lot of differentiation in how we handle maximum availability, eliminating both planned and unplanned downtime, how we protect information in the system to give you great data security, and what we're doing with our management packs, particularly with the new introduction of Oracle Enterprise Manager 12 at OpenWorld, which is a huge new release of Enterprise Manager. It's the largest single release we've done in five years, and it has a lot of new capabilities that allow us to get more management solutions sold into our install base of database and middleware customers.
In a nutshell, the heart of what we do in the database product line, in our middleware product line and applications is we're looking three, four, five years ahead and building in some fundamental technology into the software stack to differentiate and fundamentally offer competitive differentiation between what we do and what other vendors do. At the core of it, to then compete with us, competitors have to build not just one feature, but they have to build all of these features together in order to deliver a product that is competitive with what we offer. That's part of the reason why in the database business, not only have we been the market leader, but we are growing share faster than all the other competitors. Now, the combination of all of these software innovations coupled with hardware is what we introduced with Oracle Exadata.
I will try to explain Oracle Exadata in as simple as possible terms so that you understand why customers are so interested in it. If you look at a traditional database running on a computer processor, whether it's a data warehouse or it's a transaction processing system, the average processor utilization on the processor nodes of the database is between 15% and 20%. The reason it's in that range is not because the customer bought five times the necessary hardware. It is that the database has traditionally been bottlenecked on some other part of the technology stack, preventing you from going any faster, even if you bought 10 times more processors. As a simple example, you've got these processors that are very, very fast, waiting for data to come on disk. They're basically idle, waiting for data to come off disk.
Even if you bought 10 times the number of processors, you could go no faster because you now have 10 times the number of processors sitting there waiting for data to come off disk. What we did with Exadata is basically solve four simple problems by doing it all together. The first part of it was something we call smart scan. What smart scan is, is basically we said, instead of having these processors wait around for data to come off disk, let's push a bunch of the query processing down to the storage layer so that some part of the query processing happens at the storage layer. It happens very quickly because it's right next to where the data sits. Then only results come off the wire into the compute processor.
By definition, you're sending a lot less data across the wire back to the compute processor, which means you're not as bottlenecked anymore. That's part one. Part two is we put, instead of having a 10 Gb Ethernet link between the compute processor and the storage area network, a very fast pipe based on InfiniBand between the compute processor and storage. We've got as much as 880 GB per second coming off the storage layer, almost 1 TB per second coming off the storage layer, which means you get very, very fast access to large volumes of data. That's part number two.
Part number three was now that you've got so much data coming off the wire so quickly, we put in a large cache inside the database, meaning the database cache, which is the memory cache of the database called the SGA of the database, can now access a lot more data in memory. The reason for that is all the data that's been coming from this very high-speed network, you want to store more of it in memory and serve it out of DRAM so you get instantaneous response. Finally, now that you've got all of that data available to the processors, we parallelize access to the data with parallel query. You've got a lot of different queries that are pulling data from this high-speed in-memory cache very, very quickly.
It's a combination of those four optimizations in the software coupled with a very high-speed networking and storage layer that gives customers huge performance speed up with Exadata. Okay, now finally, all of that is transparent to the application. You do not have to rewrite your application to get these performance benefits. Any Oracle Database 11G application gets the performance benefits of Exadata without requiring application code change. The reason that's helpful is it allows us to appeal to a very broad set of customers and a huge install base of customers with these products. Now, what I'd like to do is invite Andy Mendelsohn from my team. Andy, come on up.
Andy runs our database team, and he's going to go through some more examples of some of the new products we introduced at Open World and also talk about what some of the customer benefits some of our real customers are seeing. Andy, you go.
Thank you. Okay, let's go to the first slide. Oh, okay. I'm going to first talk about in-memory database. There's a lot of noise out there from SAP talking about in-memory database. You think the whole objective of the world is doing in-memory database. What customers really want is they want really high performance, and they want it at lowest cost. They don't want highest performance at incredibly high cost because we can always figure out how to do that. The real market is high performance, low cost. That's our goal. It's a big difference between what we're doing and what SAP is doing, just looking at that goal. If you look at this table that I have in the chart here, we use Exadata, some of the speeds and feeds in Exadata to sort of explain what we mean by highest performance at lowest cost.
If you look at this, you can see that the cost of storing data on disk is one-tenth what it costs on flash, which is one-tenth what it costs in DRAM. Now, the cost per terabyte, there's a 100 : 1 difference between the cost of DRAM and the cost of disk, right? If you want to have big data in your database, you know, we're talking now petabytes of data, it's prohibitive. The cost of storing all the data in DRAM is just way too expensive for customers to afford, except, you know, a couple of customers in some niche industry might be willing to pay the price for that. When SAP talks about their in-memory database, they're assuming all the data in the database is in memory. Number one, they're just pricing themselves out of the market from day one with that strategy.
What we're doing is quite different. Our idea is we want to use a memory hierarchy or storage hierarchy for storing all the data. All the data in the database, you've got petabytes of data, it all goes on a hard disk at very, very low cost per terabyte. With our compression that Thomas talked about, it's even lower cost. We have incredibly low cost per terabyte if you look at Exadata's cost versus storing data anywhere else, just because of our compression and the fact we use commodity disk drives for storing the data. On Exadata, the cost per terabyte is dominated by the cost of storing data on disk. We have smart caching layers that move data off disk into flash as the data gets accessed, then it gets warm. We say we have the warm data in the flash.
Then as the data is really hot and it's being constantly accessed, we move that data off flash into main memory. What you get here is the best of all possible worlds. When the data is in memory, we can scan it at incredibly high performance. If you look at the cost of storing all that data, it's all on disk. The cost per terabyte is dominated by the cost of disk, and we can deliver incredibly high performance at low cost. That's the first thing to really notice here. The next thing is, there's a lot of excitement about in-memory database, and you can see why in this table, right? On Exadata, on a single rack of Exadata, we can scan data on disk at 25 GB per second.
That is meaningless to you guys, but just as a reference point, the biggest storage arrays out there from EMC, Hitachi, they do 10 GB per second, right? We're already at 25. Of course, we scale out. If you want 50 GB per second, you just have another Exadata rack standing next to the first one, and you get a third, fourth, whatever, which is another advantage we have that you can't get from conventional storage. You move it into flash, you can go a little faster. We can go about 50 GB per second. In fact, when we have data in flash and on disk, we can scan both in parallel, and we can actually get an aggregate of like 75 GB per second, the sum of the two. As you move it into memory, we can go even faster.
We have a number here about 200 GB per second, which is about a factor of three. The other thing to understand is scanning in memory is really good, but this is not a factor of 100. It's a factor of about three, four, you can get getting the data into memory versus using flash or disk. That's really good, and we're going after that, but it's not 100 times. The final thing to note is that because of our compression on Exadata, we can compress data about 10 times pretty routinely for a BI workload and even higher for archival purposes. We can scan data. There's another factor of 10 here. Larry was really talking a lot about that on his keynote on Sunday. Instead of 200 GB per second in memory, we can effectively do almost 2 TB per second in memory.
These are incredible, amazing rates if you think about it. Our strategy is differentiated. We're using a storage hierarchy. We're trying to get highest performance at lowest cost. The other thing that's quite different about us versus like SAP is we've been in this business of running all kinds of database workloads, OLTP, data warehousing, all using the same database, same database technology. You don't need to buy Teradata to do data warehouse and DB2 for OLTP. We do everything in our Exadata appliance or on other platforms, right? SAP has from scratch now starting to build out their vision of supporting OLTP and BI someday in their in-memory database. They're starting from ground zero, you know, a couple of years ago. We've been at this for 25 years. It's taken a long time to build highly available, scalable OLTP systems, highly secure systems, highly, you know.
What SAP has decided to do apparently is they're just focusing on BI. If you look at what they're talking about, it's all BI because the task of building up the high availability technologies that we call MAA, Maximum Availability Architecture, Data Guard, Rack Clusters, you know, failover technologies is really daunting. It's going to take them, you know, it took us 20 years. Maybe they can do it in 10. That would be pretty impressive. What they're doing is not trying to do what they said their vision is. They're just focusing on BI, which is there are a lot of players out there in the BI space. You know, there's Teradata and Netezza, etc. They only do BI. That's really what SAP is up to with their in-memory technology. It's very focused on that segment of the market.
Again, we're very different with our Exadata strategy, our database strategy. We have always been focused on doing both in one database. If you run Fusion Apps, eBusiness Suite with Oracle, you can do some BI in that same database. You're not forced to move data out of your database in order to do some analytics. That's another big difference. I think I'll leave it at that. I think that's the key thing to understand is, you know, we've been at this a long time. We've been working in in-memory database technologies very long. You know, Thomas mentioned about TimesTen. TimesTen is the leading in-memory database out there. They've been in the transaction processing space for years and years. We recently have enhanced that product to do advanced analytics. That's part of our Exalytics engine. We know this space really well.
The core Oracle Database has in-memory parallel query technology that came in in the 11G release. We're working very aggressively in this space, and we're very confident that we're going to deliver on the vision of HANA, you know, real soon. We're very close to already what they're going to take 10 years at least to do. Okay. Next, what's new with Exadata? Thomas did a really good explanation of why Exadata is really fast. One interesting aspect of Exadata is that out of the box, we have this standard configuration of database servers and storage servers in the box that works pretty well across a wide variety of use cases. However, there are some customers who have a need for larger amounts of storage than what comes with your Exadata box.
They have asked us to just have basically a box of storage that they can use for things like backups, retention purposes. There's a lot of archival data you have to retain for years and years. They have asked us, just give us a whole bunch of storage that we can connect to our Exadata boxes for storing this large amount of data, right? That's what the Exadata storage expansion rack is. It's just a rack full of 18 of what we call our Exadata storage servers. It supports, it runs these high-capacity 2 TB disk drives, a total of 432 TB of storage there. We have not compromised in any way on the architecture of our storage server. It's got full, all the flash technologies in there, all the processing cores there, smart scan, hybrid columnar compression, the whole works.
We expect customers, a lot of customers, especially doing transaction processing, have this need for archival retention of large amounts of data. We think this will sell very nicely in that space. It comes in a full rack, half rack, quarter rack configuration. Big data first. Every, you know, next day in memory, the other, another big buzzword everybody wants to know about is big data. On Monday, I did a general session that laid out our big data strategy, and I announced a number of software offerings and then a new engineered appliance called the Big Data Appliance as a way of deploying our software offerings. I have to tell you, I've done a lot of sessions at Oracle, but this has been the most well-received announcement I think we've had in probably since Exadata.
The customers who really know this space, the Hadoop space, and I'll talk a bit about, are very excited about our software. You know, they already have hardware, but they're really excited about being able to load data at really high performance from Hadoop into an Oracle database with our Oracle Loader for Hadoop. The customers who don't understand this space, and there's a lot of them, let me tell you, our enterprise customers really are new to this space. They are really excited about the fact that we're going to give them a whole solution with our Big Data Appliance, Exadata, and our Exalytics analytics engine. They can very easily deploy this kind of solution. Maybe it's worth going a little bit into what this whole space is so you can understand a little bit of the technical details below all the buzzwords and noise and everything.
Let's go through some of the things we announced. First of all, let's talk about big data a little bit, first of all. What is big data, really? You know, it sounds like it's big. It's high volume, large amount of data, right? Everybody agrees on that. There's also this thing about velocity. There's usually large amounts of data streaming into your IT shop. What are some examples of that? I think the easiest one to understand I like is in healthcare. You know, you're out there, you're playing tennis or something, and suddenly you've got this little pain somewhere and, you know, don't feel real good. You go to the doctor the next day and you go on the stress test monitor and they can't, they don't see anything.
In the past, what happened would, you know, the next week, you might have a heart attack and something, you know, end up in the hospital. What they're doing today is they're putting a heart monitor on their patients and saying, okay, just wear this for a week and we'll look at the data. They either capture the data on your, you know, person or they send.
Data over the web to their computer systems, and they monitor you. Most of that data is completely useless. It just says you're perfectly fine and you're healthy. You go play tennis again or whatever, and suddenly you start not feeling well. That data is really interesting. That's what the doctors really want to look at. This is one of the characteristics of big data. There's a lot of this data coming through, and most of it is completely useless. You want to just get rid of it real fast. It's only the real interesting events, like you may be having a heart event of some sort, that the doctor really wants to capture. In fact, generally move into his relational database and analyze that data. That's a key characteristic of big data. There's a lot of it, a large volume of it. Most of it is not interesting.
We call it low information density data. There are these nuggets of information that are really important. You want to load that data into your data warehouse and integrate it with the standard data sources you've been using for years, like the data from your sales system and ERP system, CRM systems, all that kind of stuff. That's how we view this whole space. There are lots of other interesting examples. People talk about Facebook feeds. That's another new kind of big data. Every company I talk to wants to put up a site on Facebook and have a little box where the users can click and say, yeah, give me all your personal information in exchange for a coupon. Everybody's doing this to learn more about their customers, to figure out how to sell them more stuff. That's another example of big data that people talk about.
That's sort of what big data is. We think high volumes of data, a lot of low information density data. The nuggets you find in that low information data you want to load into your data warehouse to analyze and integrate with your standard BI/DW kind of process. That's how we view the whole workflow of big data. What are the products we came out with? Oracle NoSQL Database, Thomas already mentioned that. It's a business we've been in for years. We have a product called BerkeleyDB, which is the most popular NoSQL engine. We are now extending that to be a distributed NoSQL engine. All that means is instead of one index, you now have potentially hundreds of indexes.
When you insert and query data, we've built a layer on top of BerkeleyDB that transparently lets you find the appropriate index that you've inserted the data into and find it. We've also added some HA capabilities. When you insert a key value pair into BerkeleyDB into the NoSQL database, we don't insert it just once. We insert three or four copies of it so you have high availability in case of any kind of loss of data. That's the big NoSQL database. It's an evolution of existing technology that's been out there for years. That's what the NoSQL community means by NoSQL database. It's a very simple access. There's NoSQL. You just can say, give me a key. Here's a key and a value. Or give a key and look up a value by key and get one or more results back.
It's a very simple programming model that's been used for years. It used to be called index sequential access methods on the mainframe 30 years ago. There's not really a new business here. It's just an evolution of an existing technology that we think a lot of customers are very interested in. On the next side of the stack here is the Hadoop stack. What is Hadoop all about? It's what's called a MapReduce engine. It's a platform for Java developers to write massively parallel algorithms on. It's something that is being used in very sort of niche markets today. If you go to like our intelligence community, they have all this data streaming in from God knows where, right, that they're looking at. They're doing this, sifting through the data, looking for interesting evidence of some conspiracy of some sort.
When they're done with that processing, they do that on Hadoop. Then they load it into their Exadata. They have batteries of Exadata database machines there. They're doing pretty much what we're talking about here. The issue with Hadoop is that you need customers like our intelligence agencies who have armies of very sophisticated programmers who know how to do these massively parallel MapReduce algorithms. The technology is going to sort of be a niche until we can sort of figure out how to break through that and have something that can raise the level of discourse a little bit. What we're doing here is we have a product called Oracle Data Integrator that we've had for years that is sort of like that. It's an ETL technology for data warehousing customers.
What it does is it lets you automate, in a visual diagramming fashion, the process of moving data from data sources into an Oracle data warehouse or another data. It actually works on third-party data warehouses. You move the data in, you transform it using high-level declarative mechanism, which is mostly SQL, actually, and you prepare the data for your data warehouse processing. We think the same paradigm actually is an interesting one for Hadoop. We have added a new knowledge module. It's called an add-in to ODI that actually understands the Hadoop file system and lets it actually move data out of Hadoop, do transformations in Hadoop, and then load that data into an Oracle data warehouse at really high performance, massively parallel processing steps using the Hadoop technology, MapReduce engine, actually, for doing that.
Oracle loader for Hadoop is something standalone that's very exciting to people that are using a combination of Hadoop and Oracle databases. We have some other tools. The bottom line is we then said, OK, some of the customers, the guys that are our Hadoop users today, will be very happy to take our software and run it on their existing hardware and make it work. We think this is going to drive a lot of business for us in the Exadata side of their shops. They have lots of hardware. They may not buy a big data appliance from us. We now created this big data appliance for the other customers who are less sophisticated. They might not have Hadoop today, and they just want a solution they can roll into their shop and get up and running in a couple of hours.
This is a great solution for the masses in our install base who want to go into the Hadoop space. They just want an appliance. They want all the software they need on that appliance pre-loaded, installed, configured, and all ready to go. That's what the big data appliance is all about. I think I've covered pretty much everything there. Let's just close a little bit on Exadata and the Exadata business. The first business we went into with Exadata is really the data warehouse part of the database business. We had initially probably 90% of the customers using Exadata were in data warehouses. You can look here on the list. There are some examples here of very well-known customers, BNP Paribas, a certain company you guys know real well. They were a good example of an Oracle existing data warehouse customer. They moved to Exadata.
Everything's running about 17 times faster. They're using this to help their analysts in their trading rooms. Other companies like P&G, big huge data warehouses on Exadata. The list goes on. Turkcell is, I like to call them out. They're a big European telco provider, another example of an existing Oracle data warehouse customer. What they're a really good example of is our hybrid columnar compression. They had a 250 TB data warehouse on Oracle using conventional storage. They moved it to Exadata. Everything ran faster, just like everybody in this space, over 10 times faster. The big thing was that 250 TB data warehouse after compression, it's now a 25 TB data warehouse. 250, 25. That's huge. Nobody else can even come remotely close to that level of compression that we have in Exadata. That's the data warehouse side.
What's really exciting, for data warehouses, we usually sell a, they'll run one huge database on an Exadata on one rack, two racks, three racks of Exadata, whatever. If we really want to get into volume, the mass market, that's really the OLTP side of the space or a mixed workload side. We now have a lot of customers running Exadata, running things like Oracle's packaged applications. The latest is they're now running SAP. We're just going live this month with our first SAP customers, SAP certified Exadata in June. That's the real big market. Now that we've proven OLTP, and some of the customers on the list are good examples of that, China Mobile is running our billing and revenue management product from our com GBU. On Exadata, we've got Commonwealth Bank is running PeopleSoft on Exadata.
There are a bunch of other customers here, Garmin, Alcoa, who are also running applications on Exadata. We get to the really big market. When customers say, OK, I'm ready to just replace my whole Oracle database infrastructure with Exadata, this is what people view as sort of consolidation or private cloud environments. This is where we can go from not just selling a couple Exadatas to selling 10, 20, 30. Customers are now starting to look at doing that. Commonwealth Bank of Australia has a complete private cloud on Exadata as their building block. They're running all kinds of applications now on Exadata. The other exciting thing is the public cloud. Larry talked about that yesterday. Oracle is using, of course, Exadata as our building block for the database side of the public cloud. SoftBank Mobile is a good example of a customer.
They're in the process of going live with a public cloud in Japan. They're using Exadata as their building block there. I think I'll turn things back over to Thomas.
Thank you, Andy.
You're welcome.
To net it out on the database side of the business, we are the market leader. The way we grow the business is threefold: getting more customers to replace other databases and having them deploy more systems to use Oracle as their database. We are doing very well in that. You can look at the market share numbers. We are growing faster than the market and faster than everyone else combined. Number two, we sell software options to these customers. The software options are driven constantly by innovation in addressing new types of problems customers have. There is a whole raft of them. As customers move to 11G, there is a very interesting set of options in storage management, storage optimization, compression, etc., which have direct relevance to the customer's bottom line. Third is around the area of engineered systems.
With Exadata, there are basically four measures of growth: growth in data warehousing, growth in OLTP, growth as a generic database infrastructure, meaning consolidate all databases on an Exadata platform. Finally, the razor and blade strategy is to sell aftermarket products like the storage expansion racks into the Exadata base. Any of the new trends that you hear about big data, in-memory, we have covered that. We are going to be way ahead of the competition delivering new solutions for that based on the announcements we have already made at OpenWorld. The second leg of the business is middleware.
On middleware, our strategy has been very clear and steady in the market. It is to improve business agility and performance for customers with an engineered middleware suite, unifying a number of pieces of middleware into a single architecture and a single integrated system, thereby lowering the cost with which people build applications, deploy them, and manage them. The pieces of our middleware suite, at the bottom, are what we call our cloud application foundation. This is all of the application server and data caching technology you need to run very, very high-performance applications.
We are the market leader in this segment with our WebLogic product, our Coherence Data Cache, and our Tuxedo TP Monitor. Above that, there are two ways you can integrate systems: data integration tools to allow you to connect different systems together and move data between them, and service integration tools and business process management to address what's the market requirement around SOA or service-oriented architecture. Connect different systems together using high-speed messaging and be able to model business process flows that span all these systems. The next piece of middleware is business intelligence or analytics. This is the ability, once you've connected all these systems together, to get data out of these systems for reporting and analysis purposes. Content management is the ability to store unstructured data like files and documents in a central system.
The user engagement is all of the technology around social networking, portals, and customer experience management that's part of our middleware. As users access these different applications that you've built on this middleware, you can use our identity management products to secure who gets access to what information. We've got a set of management tools that allow us to manage the applications and the plumbing on which they run. The first important thing about our middleware business is that we offer a very broad suite of middleware that technically works together. We're a leader in virtually every segment of middleware. In fact, we are a leader in every single market category in which we have an offering. No one else has that breadth of middleware. No one else has integrated the pieces. No one else has the best-of-breed pieces. Why is that relevant?
No matter who our sales guys call on from middleware, there's always one piece of middleware that the customer doesn't have. If you have an app server from someone else, you don't have data integration tools. If you've got data integration tools, you probably don't use BI. If you use analytics, you probably don't have a place to put the documents like reports that come out of the analytics. The first thing is it gives us enormous scale and leverage in taking solutions to the customer base. Second, because we've engineered these pieces to work together, a lot of customers start with one piece of middleware and then pick up two or more pieces of middleware. In fact, 84% of our middleware customers today use two or more pieces of our middleware suite. That is by far the highest in the industry.
Now, what are the areas we're investing in in middleware? The first is basically as applications move to the cloud, applications sort of went client server to web and web to cloud. There's a number of features we've put into our application server, WebLogic, our Coherence Data Grid, and our real-time event processing solutions to provide a very highly scalable infrastructure for cloud-based applications. Data integration, we are the market leader with a product called GoldenGate to move data high speed from a transaction system to offload it into a reporting instance or data mart. We have the market-leading data movement technology with our GoldenGate product as part of middleware. Service integration, this is the notion of being able to connect different systems together using workflow and messaging. It's a market we entered in 2005.
We are now the market leader in that segment as measured by Gartner as well in what's traditionally in the market you guys hear a lot about called SOAR. Analytics, what we have done with business intelligence is to provide a single stack. It's a single tool set that allows you to do all the different kinds of querying analysis you want to do. Traditionally, if you have followed business intelligence, they've talked about roll-app, mole-app, whole-app, querying analysis. They're all buzzwords. What they really mean is people want to do different styles of analysis. Somebody wants to sit in front of a dashboard and go against a transaction system. They say, I don't have all the data I need in the transaction system because I don't have historicals. I've gone to go against the data warehouse. They say, the data in the warehouse I want to aggregate.
I put it in the OLAP engine. I want to access it from that system. I want to then build a key performance indicator that defines what's the business objective I want. I want to measure myself against that. A lot of competitors have separate tools for all of this. They have a scorecarding tool. They have a roll-app tool. They have a mole-app tool. They have a reporting tool. The problem that customers have is every time you bring in yet another BI tool, you have to reconcile the data in that BI tool with all the other reports and dashboards people are using within the company. Because you get different numbers, hey, he said revenue is going up. You're saying it's going down. It's because you defined revenue differently in the two tools.
Our view has been, and we spent a lot of time engineering this over five years, to get to a single unified business intelligence foundation that services all of these needs. For document management and social computing, we see a lot of interest based on both the announcement yesterday about social network. It is basically an area where a lot of our customers want to make collaboration within the organization work in a different way. The fundamental thing is the following: all of a company's information and all of the users who access that information access it through the web. Why would you force them, in order to collaborate, to take the information off the website in which they're delivering it, cut, copy, and paste it in something like email and distribute it in an email?
Social computing is really about doing for enterprises what Facebook did for consumers, which is bringing collaboration technology to the web base using the web as a medium of collaboration because all your information is on it and all your users are on it. That's got a lot of interest in our customer base. Now, from an engineered system point of view, what we built for middleware is a product called Exalogic. I'll explain at the high level what it does, similar to what we did with Exadata. With Exalogic, it is a collection of x86 servers, high-speed networking, and file-oriented storage. In Exadata, we have block-oriented storage. In Exalogic, we have file-oriented storage because applications in their middle tier read and write files. At the heart of it, it is a very similar concept.
If you look at a Java virtual machine that's running on an x86 processor, the speed at which your application can run is gated strictly by how fast the Java application gets access to data within the Java virtual machine's memory. What we did with Exalogic to break through fundamentally the performance bottlenecks is to provide a very, very highly efficient infrastructure called Exabus, which allows us to do direct access shared memory across a cluster of machines. In other words, if I have some data in my memory and you want access to it, you get it basically incredibly fast without going through a whole networking layer, part one. It's the similar concept. Bring the data to the application incredibly fast so it's not waiting around for data. Number two, allow you to address a large pool of memory through this product called the Coherence Data Grid.
We have the ability to put a terabyte of memory and to partition that so that applications get access to Java state right in memory. You can actually pull data into memory and access it directly from memory. You can parallelize access to it so that you can do parallel reads and parallel writes to that data. Finally, if you happen to use Exadata on the back end, you also get some optimizations to go back to Exadata. It's a very similar concept. People see huge speedups to Java applications, to C, C++ applications, to any packaged application, including Oracle's. I'll show you some of the data on the Oracle applications and SAP, which is also certified Exalogic. I'll ask just to give you some examples. Here are some measurements. In gray is the fastest commercially available x86 hardware, the fastest commercially available x86 hardware.
The measurements we've done on web application, HTTP response, Java applications on Exalogic versus them. It's 10 times in certain cases and 5 times in other cases. You're talking 500%, not 50%. Similarly, if people say, you're talking about scalability, what about response time? If my application is a trading desk application and I care more about how fast the UI comes back to me, here are some examples of response time. If you look at C, C++ and Tuxedo, four times faster. If you look at a business process engine, it goes nine times faster in terms of response time. In general, the simple message around Exalogic is basically you can run any application, Oracle, packaged, or even SAP, or bespoke applications written in Java, C, or C++. By definition, it applies to a huge percentage of the applications in the world.
The only one I didn't mention is COBOL. It applies to a huge percentage of applications in the world. It gives you very fast scalability improvements and response time improvements with zero application code change. I'd like now to invite Hasan Rizvi to come up and talk through this and a couple more examples of what we've done technically with Exalogic. Hasan?
Thank you, Thomas. I think Andy kind of covered some of this as well with Exadata. If you look at the parts that we're using, Exadata and Exalogic are actually built on the same platform. We look at the hardware platform in terms of InfiniBand, flash memory, storage, a lot of common componentry between them. Again, looking at the cost equation, if you look at applications, they need to access information from these multiple layers, from the memory in the server to disk. The same advantages that you see in Exadata are being leveraged in Exalogic as well. You heard Thomas talk about Exabus. I just want to give you a little bit more color. One of the questions we get asked a lot when we talk about Exalogic and these 5x, 10x multiples is, these are OLTP applications.
How is it possible when you take the same basics, the same parts that we use, and get this much performance benefit? I just want to highlight a couple of examples and give you some numbers in terms of how that's possible. Thomas mentioned Exabus, which is a messaging layer. If you look at an application, a lot of time is spent in parts of the application talking to each other. A database application spends a lot of time getting data in and out of the storage. You saw how in Exadata we've kind of optimized that. If you look at an OLTP application, it's sending messages back and forth to each other. If you use standard networking, there's a lot of overhead in that process. You've got a lot of copies that have to be made multiple times. The operating system gets involved.
We chose InfiniBand as a mechanism for the networking layer in both Exadata and Exalogic, not just because it's much faster. It's four times faster than 10 Gb Ethernet, but because it allows us certain other characteristics where we can now send messages between applications directly without going through all this overhead and all these layers. Just to put it in perspective, with standard Ethernet, you can get a message back and forth with an acknowledgment in about 250 microseconds. If you have a really fast Ethernet, 10 Gb Ethernet, you might be able to do it in 75 microseconds. With InfiniBand, we can do this in 5 microseconds, get a message between two parts of the application directly. These are the kinds of numbers we're seeing, and not just because InfiniBand is faster, but because it has a different protocol in terms of how it operates.
Another example that I want to talk briefly about is virtualization. With virtualization, of course, we have Oracle VM. Edward will talk about that. What virtualization allows you to do is to really provide isolation between different workloads. You heard Larry talk about it yesterday as well. You can now have consolidation. If you look at the middle tier platforms in data centers, there is still a lot of sprawl, as I would call it. There are lots of different applications with different servers, different architectures. There is a big demand from our customers in terms of being able to consolidate those workloads, both in terms of the hardware platforms and in terms of the platforms from a software perspective.
With virtualization and Exalogic, which is a lot of compute capacity, we can now provide isolation, being able to consolidate workloads with safe consolidation, as I would call it, without one part of the application interfering with the other part. The Oracle VM technology is the best performance on x86. Virtualization adds overhead, typically. It's got the least overhead. More importantly, it's engineered with the Oracle products and with Exalogic. We spend, Edward's team, our team spend, some of our best engineers are working to make these technologies work together and take advantage of things like InfiniBand. This is not something that a customer can easily just put together. The smarts are being built into all of our software products.
This is a great example where we have high-performance virtualization that really enables our customers now to start consolidating workloads and being able to run multiple workloads, but with the isolation that lets them not have the problem of one workload overcoming the other one. That's the second one. The third thing that I wanted to talk about was how Exalogic and Exadata work together. I mentioned how they are engineered on the same platform, InfiniBand, as a networking mechanism between the two. Beyond just that, I mentioned applications spend a lot of time talking from one part to the other. They also spend a lot of time talking to the database. Most applications will have a database behind them. They'll be sending messages. With InfiniBand, again, we have now a much faster way for applications to talk to database. The combination of Exadata and Exalogic is actually very compelling.
I'll talk about a few use cases. I think Thomas has some more details ahead. We can do that communication much faster. We've also integrated the high availability mechanisms. The database has clustering built in with Real Application Clusters that we've done for many years. In the middle tier, we have WebLogic clustering, which is clustering, again, to get high availability. We've done the work to make these two mechanisms work much better so that in the normal operations where failures, clustering is really needed for failure. You have multiple copies of things so that if one copy or one server fails, you can still send requests to another server. Clustering allows you to do that. We've integrated these so that in the normal case, we have, as Thomas referred to, transaction affinity, which is talk to the same server so you can get the best performance. Failures do happen.
We want to be able to deal with them very quickly. We've integrated the clustering mechanism so that when something goes wrong, we get notified very, very quickly so we can take corrective action that much faster. From a high availability perspective, another thing that I didn't mention on an earlier slide in terms of Exabus, because of this ability to communicate much faster between parts of our application, the other reason, like I said, you do this is for high availability, make multiple copies of everything. There is a cost associated with high availability, typically, because now you're making these copies. With Exabus, now we can pretty much make that high availability available for free. That overhead that customers are used to paying to get that high availability is no longer there because we can do this, like I said, 5 microseconds versus 250 microseconds.
These are fairly dramatic differences in terms of how people do it today. If you look at the customers, I just want to focus on two or three use cases that we see. Java performance, applications that are running in Java that need more headroom. You'd be surprised at the number of customers we have who have basically chosen Exalogic to get beyond the limits that they were hitting with their existing software. Maybe one comment before I get to that, the adoption of Exalogic has been pretty broad across all the different industry segments, across all the geographies. In the year, less than a year that we've been doing Exalogic, we've actually seen a lot of customer adoption already. Java performance, I was telling you about, Deutsche Post is a great example.
In Deutsche Post, building a new system, e-letter, they want to be able to deliver letters electronically, but certified by Deutsche Post so that I know that this letter is coming from this person. They know when I've received it. Obviously, there's less folks putting stamps and putting stuff in mailboxes. They want to be able to provide an electronic system with the validation and certification of Deutsche Post. They had benchmarks that they wanted to run. They were not able to do that on any of the systems that they tried. They gave us the benchmark. We had a two-week POC. We were able to do, and being conservative here, 5x better than what their KPIs were on this platform. Java performance, just pure companies looking for better performance, better response time. Another example is PHH Mortgage, our applications. They use eBusiness Suite.
They use PeopleSoft, two to three times better performance in terms of running our applications. Square Two Financials, a custom app, a bespoke application built on our stack, so ADF, which is our development tools, 5 to 10 times more users because of the memory advantages that I talked about and 10 times faster response time. Two other things to highlight here, Banco de Chile, they were actually with us on stage on Monday. The smarts on Exalogic, Exadata is all built into the system. The applications don't have to change. These are all standard applications. This is the beauty of it. You can take an existing application and move it and see these benefits. Banco de Chile, they were on stage with me. They talked about taking an existing application, moving it on in less than two weeks. It's not something that has to be changed in the application.
You're not starting over again. It's standard applications, standard Java programs that you can use to move over. The final one I want to highlight is Tuxedo. Thomas, again, talked about Tuxedo. Tuxedo is a C, C++, I guess, quote-unquote, "old style" application server. We've done a lot of work in Tuxedo to make it an attractive platform to take mainframe applications and move it because we've got emulation for KICKS and IMS and to make it really easy to replatform mainframe applications without having to rewrite them. Of course, there are some people who are rewriting them. It's much more efficient, in some cases, to just replatform it and get out of those millions of dollars of MIPS bills that our customers have.
Tuxedo, together with Exalogic, as a package, is now very, very compelling for those customers because now you've got a platform that can take mainframe applications with minimal changes and run it on a platform that is significantly high performance, rivaling or actually, in some cases, way more powerful than mainframe. These are some of the use cases that we've seen in terms of adoption. As we look forward, momentum is building pretty rapidly in terms of our customers. One last thing to mention, most of these customers are actually Exadata customers as well. That's an important point to highlight, which is as we talk to customers who are already using Exadata, they've already seen the benefits of engineered systems. The conversations are much easier to say, hey, we've got this. Look what Exadata did for your database. Exalogic can do that for your applications.
In fact, and no surprise, a lot of our early customers are actually Exadata customers who are seeing the benefit, both in terms of Exalogic itself, but the combination of Exadata and Exalogic working together, which actually takes the same kind of performance improvements that I talked about. You go even a level up once you combine the two. While we have a lot of Exalogic customers who are not using Exadata, a lot of these are. I think that's an interesting thing for us to follow in terms of once this message gets adopted. In some cases, the lines of business are so the requirements they have can only be met with Exalogic. Some of these are driven by lines of business. Some of it is driven by IT.
Overall, for the 9 to 12 months that we've been doing Exalogic, it's really been a compelling value proposition for our customers. I think that might be my last line.
Thank you. Thank you, Hasan.
Sorry.
Exalogic, a middleware business, just to net it out, best-of-breed pieces and a large number of them. This means no customer has all of the middleware and allows us to enter as many customers as possible. Number two, because it's engineered to work together, we upsell and cross-sell pieces of middleware because it's easy for a customer to say, if I'm going to get you analytics, I might as well use your document management. If I'm going to get your document management, I might as well use your security stuff. That's what's driven our middleware business over the years to be the market leader it is and continues the growth. Third, Exalogic allows us to offer a unique value proposition in middleware. No one else offers an engineered system for middleware. The performance advantages are so significant, it does become a competitive differentiator against other players like IBM.
Finally, for transaction processing customers, we get a two, let's call it a, we can sell Exadata for the database and we can sell Exalogic for the middle tier. You can sell two engineered systems into the same environment that a customer is operating. The other system we announced this week was Oracle Exalytics. What Oracle Exalytics is is a middle-tier engineered system for the person who runs a data warehouse or wants to put an analytic tool or a reporting tool in front of it. What is Oracle Exalytics? It is a system designed to give customers the fastest roll-up, mole-up business intelligence performance. When I say roll-up, mole-up, traditionally called querying analysis is what we refer to as roll-up or analysis in front of a data warehouse. Mole-up is traditionally what has been in the industry called online analytical processing. It's the fastest system for those.
It is also the fastest system for financial and operational planning. If you're in the budgeting office of a CFO, we have got a set of applications, our Hyperion applications, which are the market leader in that space. It makes them run incredibly fast and also gives you the best cost performance. How fast does it really go? Here are some numbers. These are both based on benchmarks we did. We also have customers who spoke at our different keynotes talking about that these numbers were actually conservative, meaning what they found was much faster than what we've actually got. On the left in gray is the fastest alternative x86 stack from a competitor. In red is basically how much faster we go.
We go 20 times better in response time, 15 times better for scalability for querying analysis, 80 times faster for reads, and 20 times faster for writes for online analytic processing, and roughly 5 times better for planning and 15 times better response time. How does Exalytics actually work? Exalytics works as follows. There are two basic things to it. If you think about a person in your organization who's sitting in front of a dashboard, they are using a pivot table. They add a new column to the pivot table. What the dashboard does is it says, got to get data to populate that column. What I'm going to do is convert that into a query. I'm going to change the shape of that query and then send it over to a database instance, get the data back.
I've got to remarshal that into the format of the table and then put it back in the UI. There is quite a bit of processing that goes on in order to do that. Our basic design principle with Exalytics is what happens if you could get all the data that users access directly from memory at the speed of DRAM? The first piece of Exalytics is an in-memory analytic data store. There are two capabilities there. There is a capability in timestamp to show results, to store what we call results sets. Results sets say if there's 100 of us accessing dashboards, we all access and typically send a shared number of queries to the back end. We're all looking at financial numbers. There is an overhead because every user is requesting the same data. We're re-executing the same query on behalf of all those users.
Instead of doing that, cache the results set, compute it once, store the results set in memory. When the system asks for it, just give the results back instantaneously. The first piece of it is an in-memory cache inside Exalytics, which is a terabyte-sized in-memory cache. You can store a huge amount of data there. The second is we have algorithms in our business intelligence tool called Oracle Business Intelligence to figure out what's the data that you should really put in the cache. The way it does it is it's got a capability that's watching all the queries going through the system. It says, hey, you know what? These queries happen all the time. This data set is being accessed all the time.
Instead of having to have a user figure out what data to put in the cache, we automatically figure it out and put it into the cache. That's part number two. Part number three is we've got something called in-memory columnar compression. In Exadata, we have something called hybrid columnar compression, which is to compress columns when they're stored on disk. In Exalytics, what we have is a capability called in-memory columnar compression. What it says is if you fetch one of those columns into memory, we'll compress it as a column vector representation in memory, thereby allowing you to take that terabyte of memory that we have and cram much more into that terabyte of memory than just what traditionally 1 TB could hold. You can actually get compression ratios of as many as six times with in-memory columnar.
All of this says the fundamental thing is we're very smart about putting as much data in memory, optimizing what data is actually brought into memory, compressing things so you can cram as much stuff in memory, and then, most importantly, parallelizing access to it so that you've got lots and lots of threads pulling data out at the speed of the processor. We built a brand new visualization interface that customers love because now you can get access to data so fast, you can do kinds of analysis that you couldn't do before. For the performance side, for the planning, what we did was something different. On the planning side, we have a planning engine called SBASE. SBASE reads and writes blocks of disk, and it uses different representations on disk if you're doing reporting on the budget versus if you're updating the budget.
What we did with SBASE was basically say, OK, that lower half of SBASE that writes blocks on disk, we're going to modify it so that it writes certain blocks directly into memory and actually doesn't persist them. If you're doing calculations all the time, you can put it into memory and get access super fast in memory. The second thing we did, that's what we call the in-memory cache underneath SBASE. We have a smart storage manager similar to the analytic layer saying, I need to figure out what queries are going to the system. There's a storage manager that says, I need to know which blocks to put into memory. We have a lock manager where we eliminated certain lock contention so you can get massively parallel scale.
You can do lots and lots of cubes and updates on those cubes and load data out of them very, very quickly. The fundamental premise inside Exalytics was the following: bring as much data into memory, be very smart about what data you put into memory, eliminate lock contention so you can do parallel read-write access, and then give you amazing performance improvement as a result of that. This product has been in early adopter phase. We have a number of beta customers. There were quite a few of them who were on stage with us talking about the performance numbers they're seeing. Frankly, their numbers are much faster than even what we're showing in comparables. What's the net-net one thing I wanted you guys to get across Exadata, Exalytics, Exalogic? If you're a customer today, why are customers?
You guys have heard a lot of buzz at OpenWorld about it. Why are customers excited about it? It's very simple. If you look at Intel processors or any other processor in the market and you look at how long it takes them to give you a five times improvement in performance, that's a generation of technology. It's typically six or seven years before you get that. You have a customer sitting here in October 2011 saying, my application is going to go five times faster in 2018 if I'm lucky. The alternative is you can buy a system. It runs standard operating system, standard Oracle Database software, standard Oracle middleware. You do not have to modify your application. Tomorrow or a couple of weeks from now, you could get that 10 times performance improvement.
It is that change, that fundamental change in performance, scalability, and reliability that our customers are very, very excited about and the fact that because you do not have to change your applications in any way at all to get that performance advantage, they don't have to go back to the developers in the organization and say, you know what? Got to recode all your crap to get it to go faster. It applies to every system in their company. Let's now talk about applications. Our strategy with business applications is to offer significant functional value to our customers with easy-to-uptake Applications Unlimited releases: PeopleSoft, Siebel, E-Business Suite, etc .
Second, to deliver Fusion applications, which are now generally available alongside these Applications Unlimited releases, and to ensure that customers can adopt Fusion applications side by side with what they have so that you do not have to rip and replace what you have in order to start using Fusion. Third is to provide specialized industry-specific line of business applications that drive business processes that are transforming in a variety of industries, whether that's the way that converged billing is required in communications or what's happening in core banking, etc. Finally, to put all of them on our common technology infrastructure. How have we done just in the last 18 months? If you look at our Applications Unlimited products in the past 18 months, here is what we've delivered in the horizontal products. In a nutshell, a ton of new stuff has been delivered. Pick an area, PeopleSoft.
PeopleSoft 9.1 and 9 are the largest releases that PeopleSoft has done in 11 years. When I say 11 years, that's a long time before even Oracle bought them. No customer of Oracle's has the competitors spend a lot of time saying if Oracle is doing Fusion, it means it's not doing the other stuff. Absolutely, positively not true. The proof of it is what we've delivered because our customers see what we're delivering both in terms of features and in terms of the new modules and how easy we have made it for them to uptake it. The same thing applies to our global business units. They've done a lot of work delivering a lot of new products in the last 12 to 18 months. Here's a list from them.
I'm not going to go through every one of these, but I wanted to give you a sense of how important some of those things that you guys delivered are. Here's an example of E-Business Suite 12.1. They span all parts of E-Business Suite from ERP, supply chain, human capital management, as well as the technology foundation. The same thing applies to Siebel. The first message I want to make sure you understand is we are getting strong business applications growth. You guys have seen our results in that space. We are capturing market share faster than the competitors in that space because customers are comfortable that, one, they're getting new functionality, and that functionality is easy to adopt.
The upgrade point I really want to point out, if you look at our applications customer base, 54% of our applications customer base is on a release that was delivered in the last, it's in the two most recent releases of that product line. It is substantially higher than anyone else. In the last 12 months, 67% of our Applications Unlimited customers have bought a module or a product from a product line that they didn't actually come into Oracle with. If you're a PeopleSoft customer, 67% of them bought Siebel, Hyperion, Agile, or some module of E-Business Suite. If you're a Siebel customer, 67% plus bought ATG for web commerce and industry application, etc . We have a huge breadth of applications, and we're able to cross-sell them because of the continued functionality delivery on those products and the fact that we're integrating them together.
Most importantly, these products also run on our engineered systems. If you look at the performance, here is a comparison of E-Business Suite, PeopleSoft, and Siebel on Exalogic and Exadata against the highest performance corresponding in x86 hardware. Frankly, the same can apply to any other hardware. This measures response time. If you look at this and go, that box, the red box is smaller. It is response time, and so smaller is actually better. There are big, big performance improvements here and also in the different industry applications, JD Edwards, etc . One accomplishment that we are incredibly proud of as a development organization is the delivery and general availability of Fusion Applications. To talk about that, I'd like to ask Steve Miranda from my team to come up. Steve, here you go.
Great. Thank you, Thomas. First, let me cover what Fusion is across a couple of different dimensions. In terms of overall breadth, Fusion Applications encompasses seven different product families and over 100 different applications. It spans everything from financial applications, so things like GL, AP, AR, human resource applications, which includes not only core HR, payroll, benefits, talent management, talent review, goal setting, includes supply chain management, largely around an application called distributed order orchestration, inventory, costing. I think you guys get the idea. Project management, which is generally used by services organizations to track costs and apply that cost to different projects. Procurement, which tightly integrates with the back office and financials. Customer relationship management, heavily focused on sales and marketing, including mobile.
Finally, governance, risk, and compliance, all of which can be adopted either as a complete suite, very similar to the architecture that the eBusiness Suite had, or what's the most popular with early adopters, to be adopted modularly in a coexistence fashion, as Thomas described earlier, which has the ability to run side by side to the existing investments, not rip and replace PeopleSoft or the eBusiness Suite or Siebel or JD Edwards. From a product footprint side, it's not only complete and broad, but each of these products is very deep. I think you'll see that when I describe the customer use cases in a moment.
Now, in terms of what it is, in addition to the kind of functionality we think differentiates within the functionality of the finance, CRM, etc., three big kind of driving tenets that we had, which we think differentiate Fusion from anything on the market. The first is it's the first application suite built on standards-based industry-standard architecture. That's things like Java, HTML, BPEL for the business process workflow, all web services built, all communications done and rendered in HTML, which also gives us the ability to run mobile. There are some interesting technical things about that. From an application perspective, it's a cost and risk perspective because any other platform, any other vendor before has built basically a proprietary platform first and then built their applications on top of that. What does that mean? To run, manage, extend, maintain Fusion Applications, you need to know Java.
If you're basically a computer science graduate from any university anywhere in the world, you know how to run, manage, maintain, extend, configure Fusion applications because it's just Java, standards-based Java. There's no learning of a force.com. There's no learning of ABAP or any other kind of proprietary standards-based language. There's a technical discussion, but it's really, from a business perspective, cost and risk. It's cost and risk now and especially going forward over time. The second big driving tenet is it's available both on-premise and in the cloud. That really is born from that standards-based architecture so that it can run in the Oracle Cloud where we support industry standards. It could also run in private clouds. If you'd like to run it, go ahead and run it on-premise.
Oh, by the way, since it's built on standards, if you have to run another application, a third-party application not built by Oracle, the chances are that it's built understanding Java, web services, rendered in HTML, running with BEPL as a business process workflow is high and growing as opposed to a proprietary platform where if you have an ABAP application or a force.com application, you have that application that has to be built on that particular platform. It's all built on industry standards, all available on-premise and in the cloud. The final thing is on the design point of how we design it in terms of user interface.
This complicates a number of different perspectives, but just to give you an idea of why this is important, where customers are seeing this pressure, in any previous application build, there was very little user interface expectations that we had in common. By we, it's all of us in this room, all of our mutual customers, their customers, their suppliers. By that, I mean the following. You had a Windows application, and we all knew what the little X meant on top of the window. You click X, and it closes it. You have the little dash, and you click the dash, and it moves it to the desktop. You have this box. When you went into a business or you used a website from or you were a customer or a supplier of a business, there was a natural training of the computer system.
It really didn't matter whether you had to learn to hit Control, Control N, or you have to hit F11, or if you have to hit backslash Z. Whatever the mood was, you went in, and you had to be trained on the computer system. Some were better than others. Fundamentally, that was the first step in training. Today, we all share much more common metaphors for transactions. We all know how we expect search to work, whether you use Yahoo, Google, or Bing. You have a common metaphor for search. We all have a metaphor for how you expect to buy things in a shopping cart metaphor, whether you use Amazon, whether you're shopping for flights on Priceline, or using eBay for auction type of interfaces. It's all kind of a common user interface.
More than that, though, we all now have an expectation of business intelligence embedded in those transactions. On Amazon, you go and you have other users' recommendations. You have charts. You have business intelligence that tells you, oh, if you want to buy this product, you might actually want this peripheral, this accessory. You go onto Netflix. You have this movie, and oh, this movie is just like that. If you liked it, you have this here, all of which is business intelligence. It's not kind of business intelligence in the way we classically thought of it with charts and graphs and trends and KPIs, but it's all business intelligence about what you need to know, what you need to do. Part of that business intelligence is the social network.
In the enterprise, the social network is more than just people you know and relate to, like on Facebook or Twitter, where you seek out friends and people you know. In addition to that, it's business transactions that you care about. If you're in sales, not only do you get the social network from people on that sales team or people you follow and you know, but you get any news from the accounts that you follow or any news from accounts which may be referenceable to the accounts that you care about because the business intelligence system knows that is a customer that you care about and relate to.
All of Oracle Fusion Applications built from the ground up had modern UI concepts with the things that I talked about before, like the search metaphor, the shopping metaphor, with embedded business intelligence and embedded, quote, at the beginning of Fusion, we called collaboration capability. Now it's really expanded to be social capability and all designed around the GUI questions to help users actually use the application efficiently. What do I need to know? What do I need to do? How do I need to get things done? Who do I need to contact? How do I reach them, the social collaborative part? All based on standards, all available in the cloud or on-premise, and all with BI and social collaboration at the heart of the user interface.
I mentioned at the beginning, it's also a consumable in what we call a coexistence fashion, which simply means you can deploy a product family, HCM or financials, the whole suite, or particular products. For example, you may be a PeopleSoft customer and deploy talent management in Fusion while you keep PeopleSoft in place. You might be an eBusiness Suite customer and deploy distributed order orchestration going forward. You may be a JD Edwards customer and deploy Fusion CRM going forward. Here are some of the popular examples. Rather than do that, let me try to call out a few customer examples. We have over 200 customers signed up with Fusion today. These are just some of them who are using the product as of now. I'll start from the back office and move forward in my examples and use cases.
There are a few customers who are doing a bit more of a complete solution around financials. This is all the financials, procurement projects. Red Robin Restaurant Group is one example. They have a JD Edwards World implementation today. This is our oldest product, AS400. They have grown substantially and really a turnaround story financially going forward. They need a newer and modern platform system to support that growth going forward. They've chosen to do the reimplementation from JD Edwards World on the Fusion applications, all in financials. They have just extended that now to include HCM because they've started. That's a hosted implementation. We have a very similar story with the PGA of America. They are doing it from an eBusiness Suite release and also with Pacific Northwest National Lab, who's upgrading a set of custom financial applications. Moving to HCM, two examples come to mind: Principal Financial Group.
The Principal Financial Group runs both finance and HR in PeopleSoft. They've just completed a PeopleSoft upgrade to 9.1. They are implementing Fusion talent management and compensation that integrates with PeopleSoft. In other words, PeopleSoft HR Pay remains on-premise. They've integrated with Fusion in the cloud to Fusion for talent management and compensation. When they do the ratings, performance appraisals, and then raises, bonuses, stock options, etc , they feed that back into PeopleSoft. They run PeopleSoft payroll on-premise as they would going forward. It's an HCM coexist. Very analogous to that, similar is Alcoa, who's doing the same thing. Only they're doing it with eBusiness Suite, and they're doing it on-premise. By the way, Alcoa is also an Exadata customer. Moving forward to the supply chain, we have a brand new product we call distributed order orchestration, which is really a service-oriented order orchestration, order management system.
Boeing is implementing distributed order orchestration on top of a number of different ERP systems, including Oracle eBusiness Suite and including SAP. In their aftermarket kind of market, they have the very predictable kind of market for planes and the builds that they do. In the aftermarket area, they have a little bit more of a dynamic order orchestration market for them. What they needed was a way to take a view across their different ERP systems and do order orchestration. They're not upgrading SAP nor eBusiness Suite, but instead implementing distributed order orchestration to take orders in, have global visibility of their order fulfillment, different mechanisms that are in their different ERP system, and use our order promising to route those as appropriate. That's a supply chain example. Moving all the way to the front office, Green Mountain Coffee Roasters. Green Mountain Coffee, I guess now midsize growing largest.
They're the second largest growing company, according to Forbes, for the last two years running. They have both PeopleSoft originally in CRM, and then through an acquisition, they have Siebel. They need a brand new CRM system. They do a lot of the sales through distributors, not through their direct employees, but through partners. They choose Fusion not only for the functional capability to do the partner relationship on the CRM side, but also to consolidate both PeopleSoft and Siebel. They're upgrading both those releases to Fusion CRM. They're doing it SaaS, and they are continuing with their PeopleSoft in the back office moving forward. What I gave you is just some examples of a variety of different customers, a variety of different industry sizes and shapes.
I talked about all the product families, financials, HCM, supply chain, CRM, some of which are running in SaaS in the public cloud like Green Mountain Coffee, some of which are running in the
Private cloud SaaS, like the PGA and also Principal Financial Group, some PeopleSoft customers, JD Edwards customers, Siebel customers, and E-Business Suite customers. Hopefully that gives you an idea of the depth and the breadth of the coverage. I think we have a video, or are we going to cut to the tip of the video?
If you don't mind, we'll play it at the end. We have a video for you guys to hear from the customers directly in case you think that you want to hear directly from the horse's mouth what they like about Fusion.
Thank you, Steve. Our basic view is as simple as this on the application side of things. You know, we have a very broad suite of applications. No one else offers the breadth we do. We allow customers to pick up individual modules, as well as to do replatforming. Fusion, which for a long time our competitors took shots at us saying, why would you want to rewrite a suite of applications? A lot of our customers are asking the competitors now, how long will it take you to rewrite your applications?
Because we've allowed customers to adopt these incrementally without doing a rip and replace, you'll see Fusion is going to be a great asset for us as we go forward against competitors who are on legacy architectures and have no plans or design to rewrite. It's fine to throw stuff at competitors saying, oh, their strategy is weird because they're doing this. If their strategy gets right, then you have an issue because now you're going to have a hard time catching up. Fusion was a six-year effort for us, and no one else is going to be able to do what we did without taking six years as well. Finally, introducing the Oracle Public Cloud, an announcement we made yesterday. What are we doing with this? It's very simple.
We're going to deliver through cloud.oracle.com, a public cloud environment where a customer of Oracle can go and get access to our database online, our WebLogic server online, can get our CRM applications online, can get our human capital management applications, talent management, social networking, and security all online. How does it work? As a user, you go to our cloud.oracle.com, you pick a service, you select the specific plan you want, you enter your credit card number, get your credentials, and you get going 20 minutes later. It's a subscription-based model. You can get access to any of these capabilities all online at cloud.oracle.com. There are some capabilities we built into it. You have the ability to provision your cloud. You can choose from a variety of different configurations for the services. For the database, you can say how much storage you want, how much data transfer you want.
You choose it yourself. Once we provision the cloud for you, meaning the instance for you, you can manage it yourself. We give you a version of Enterprise Manager called Cloud Manager, where you can manage it yourself. If you want to figure out service levels, how much downtime have I had, how many users of mine are accessing the system, how many transactions went through the system, you don't have to call Oracle. Self-service monitoring, we give you a simple interface where you can go and get all the SLAs about your service, cloud.oracle.com. Suppose we give you a CRM application. You said, I want CRM, I want access to it. How do you configure it? By configuring, I mean, how do you define how many salespeople I have, what territories do they belong to, how do I map accounts to those territories, etc ?
All of that is done online through a browser. You go in, we ask you a set of yes/no questions, we configure the system for you. Value, a business user can use the cloud without any developer or IT involvement. If you want to extend the cloud, you say, you know what, I really would like your order definition has 13 columns or 13 attributes. I want to add four of them of my own that represent my orders, which are a little bit more complicated than yours. How do you do that? No power tools, no IDE, no nothing. Go to the browser, you can extend our data model, business process, business logic, reports, etc . All of it can be done through the browser by a business user without any coding.
What we're basically doing is taking the power of our software and making it available instantaneously through cloud.oracle.com to any customer of Oracle's through a subscription. How is it built? What's unique about it is not just that we give you the breadth of this software all integrated, but also the way it's architected. Our cloud runs on Oracle Exadata and Oracle Exalogic. All the databases are isolated database instances, isolated at a physical schema level, not multi-tenants from the point of view of data striping within a database, very important to our customers on Exadata. All the middle tiers run with virtualized instances using the virtualization capabilities that we offer as part of Exalogic. You get elastic capacity. You say, you know what, this next week I want a little bit more capacity. You can self-service request it, we'll scale it up and down for you.
If you get to a point where you need instantaneous scalability, which is as the workload runs, we can also do that for you. Security, you know the data is 100% isolated at the disk, at the table space, in the database, on the wire, in the application, out on the network to the user. For every tenant, it's 100% isolated from every other tenant. Finally, we give you tools to integrate our cloud with your enterprise. Obviously, what attracts people to our cloud is what it runs. We run the world's leading database and the world's leading application server. Why is that relevant? People can use our cloud to run the applications they use to build for the Oracle Database, not for a tiny niche database which has different SQL dialect, etc , that nobody cares about.
We're the only cloud that runs standard Oracle Database applications and standard Java EE applications. Point number one, so it applies to everybody. Point number two, we obviously feel very strongly we've got great applications. CRM, a full suite available as part of our Fusion Suite. By the way, if your salespeople are good guys, but you want to do a talent review about those sales guys, you know we give you HR in our cloud. You don't have to say our cloud only runs CRM. You want HR, take the data out of our cloud, put it in somebody else's cloud, and that's your problem. We don't do that. We give you CRM, HR, database, Java, talent management, all in one cloud, all integrated. You don't have to move anything around on your own. Finally, we also introduced a new product called Social Network.
The basic idea is that in addition to giving you the access to this cloud, we also allow you to collaborate through the cloud with all the people you do business with within your company, salespeople talking to legal people, salespeople talking to HR people and contracts department, as well as with partners and customers outside your company, all in one place. Finally, there are two other aspects to our cloud. We believe that the value of a cloud offering improves the further up in the stack you go. In addition to offering you just a basic HR application or a basic sales application, which in our view, our offerings are highly differentiated, we're also going to give you content with those applications. Let me give you two examples. You're a small division of a company.
Today, you can buy our CRM applications or those from other vendors and use them to manage your pipeline. However, before you build a pipeline, you've got to get customers. There's a process that people go through in buying leads lists. They go to a company and buy a leads list, which is a name of a list of companies and who in that company you might call on. Many times, that's an old list and the customer, the person has moved on. You've got to go to LinkedIn and figure out if that person is still doing that job or not. You have to get that person's exact email ID to send them an email saying, can I talk to you? That whole process of buying leads, cleansing them, standardizing them, and then delivering them is today a job that's pushed to the customer.
An example of content that we're going to offer is we will provide you through our applications the ability to get this information delivered to you. HR, benefits. If you're a small company, if you're a big company, you already know what benefits providers are there because everybody's calling you. If you're a small company, you're trying to figure out there's 100 benefits providers, which one has what offering? We will deliver that through the cloud. We connect you to all the benefits providers and provide an interface through which you can compare all the benefits of products and pick the one you want. Differentiation by moving up the stack. Finally, we will run not just our applications, but any third-party application that connects to our Fusion SaaS offerings, our database cloud, our Java cloud. Why a customer is very interested in this?
It's the only cloud with a complete suite of SaaS applications. HR, CRM, talent, etc., etc., etc. We are the only ones that offer all of them, which eliminates the need for them to have fragmented data across 16 different clouds. It's the only application, only cloud that runs any application without change. Database or Java, running on-premise in the cloud, you don't have to modify your applications. You just have to change a single configuration parameter and it runs in our cloud. It's the only cloud which allows customers to have the choice of running in our cloud today and back on their premise someday in the future. Why is that relevant? The choice of whether you go SaaS or whether you go on-premise for many customers is a question of how they're managing operating expense versus capital expense. That changes over time for different customers.
Because it's the same code base that runs in the cloud and on-premise, you can start in our cloud and one day in the future tell us you want to take it back on-premise. Of course, you can take it. In closing, what have we been doing in our software business? To net it out, there are basically four important things. We have a leading, rapidly evolving, incredibly broad, and best-in-class software portfolio, highly differentiated technology and business applications. Second, we have a highly differentiated engineered systems portfolio, which certainly helps our hardware side of the business, but equally importantly, gives us on the software side a competitive advantage against all pure-play software vendors.
Third, we are introducing a highly differentiated public cloud offering, differentiated in four important ways: the breadth and depth of the services, the kind of capabilities to allow you to use applications on the cloud or on-premise without change, the ability to move any application on the cloud or off without change. Finally, highly differentiated content, where the further up the stack you go in the cloud, the more differentiated you get. It's the combination of these factors that positions us incredibly well for continued strong software growth. With that, I'll take a few questions.
Thomas, just two clarifications. One on the public cloud. What's your general availability date for that?
I'm going to tell you very, very soon, a matter of weeks.
OK. On Exalytics, can you just give us a sense of where that falls in on a price point relative to Exadata and your new database appliance?
We haven't publicly disclosed pricing on Exalytics, but I will let you do the math. The simple thing is that it's a four-socket x86 server with 1 TB of memory. You can do the math on how much that roughly will cost. On the software side, a customer who is an Oracle Business Intelligence customer doesn't have to buy any new business intelligence software. It's the same software. It just runs a lot faster. If you want to use TimesTen, there's a TimesTen SKU, and that's available now, and it's about $20,000 a processor.
OK.
All right.
OK. Next question. Yes, Kier.
Yeah, hi, Carl Kirst at BMO Capital Markets. I had a question about Fusion Applications. It seems to me that if that's really going to get traction this fiscal year, you need to get some success stories on the CRM front in particular. I'm guessing part of the reason we're hearing a lot of Salesforce.com criticism from the Oracle leadership team is because you're gunning for the CRM space. What I'd love to ask you is, how does Fusion CRM differ in terms of functionality from Siebel On Demand? If that's not really the space you're going to go after, how in terms of feature functionality is it superior to Salesforce.com?
I'm not going to comment about why there's been back and forth or anything like that. I'm just going to talk about the functionality differences. First of all, we have a large number of customers in our early adopter program who are on the slide, as well as referenceable, who are talking about CRM. If you're interested, we can point you to information about that. The functional differences, if you look at what we have in Fusion CRM that attracts a number of customers, in fact, virtually every customer who is part of our CRM early adopter program has considered the competition that you mentioned because it's available today. It's a choice. Now, what is the modules? If you think about sales, the first step you go is you decide, I need a list of accounts, and I need to get master data about customers. A single place, a customer.
I need to get a leads list. I need contacts. I need to plan my territories. Right there, from a customer master point of view, we have a customer master. The competition does not. For territory planning, if you want to do anything more, like you want to plan how salespeople map to accounts and territories, are they going industries or geographical? Do you have global cover, global accounts, and not global accounts? You can, with Fusion, plan sales territories in the system. You don't have to go anywhere else to do it. The competition does not have that. If you want to do quota and commissions planning, we provide that in the box, which is typically the next phase that you do. You map all your guys, and then you decide where you're going to assign them and what quota plan and commissions plan.
The competition requires you to use a third-party product called Kaleidos. We provide that in the box. If you then say, OK, great, I have a list of accounts. I want to figure out what products to sell those accounts. We have a tool that allows you to identify white space in the accounts called Sales Predictor and White Space Analysis and give you a list of potential product opportunities to go after. Competition does not have that either. You go to forecasting, pipeline management, and campaigns, which is the next set of things you do. You say, I got all my accounts. I know what I'm going to do going after them. For campaign management, we both have multi-channel campaign management. We both have Google AdWords. We both have email marketing.
What we have and they don't is the ability for a salesperson to create his own personal sales campaign, which a lot of salespeople want to do because they don't want just corporate driving all the marketing and lead generation programs. Finally, to complete the cycle, with regard to incentive compensation, I did really well. I need to be paid a certain amount of money. Again, we give you a solution in the cloud so that you can use it. Competition does not. Even if you leave aside all the other capabilities that we offer, integrated business intelligence, social networking, etc , just as a pure-play CRM application ready for customers to use on a SaaS basis, we're very confident functionally that we have a broader, better integrated footprint than what competitors offer. Next question.
Hey, Thomas, it's Adam Jones from Morgan Stanley. A lot of the partners we spoke to over the last few days talked about HCM as an area where they're seeing particular traction. Is that by design, or did we just talk to a small subset that happened to see that as an area of strength? Why would that be?
If you look at the broad suite as four buckets, ERP, which is financials, procurement, project management, and supply chain, human capital management. You say ERP, HCM, CRM, it's roughly in our actual early adopter program, they're roughly a third, a third, a third. Maybe those partners had built a practice. Different partners build practices around specialization they're going after. It may be that they picked HCM as a very interesting area for them to go after.
Just a follow-up to that would be, we had a lot of discussions about when you get to the full suite being available in Oracle Fusion Applications. Maybe walk us through what the roadmap is to a full E-Business Suite equivalent being in Fusion.
Yeah, the primary piece, and I'm not going to give dates, but the primary piece that we will have in a follow-on release, what's the single piece that's not there in Oracle Fusion Applications right now is supply chain for manufacturing companies. We have supply chain for retailers, for telecom companies, etc . We don't have that piece for manufacturers. That's in a follow-on release called, let's say for round numbers, the next release of Oracle Fusion Applications. Maybe one last question. Yes?
Hi, thank you, Thomas. [crosstalk] of Merrill Lynch. As you build out the Fusion suite in the cloud, and it looks to be functionally very rich, do you get to a point where how do you get over customer potentially not sure which they should choose, the on-demand model or the on-premise model? Do I do the subscription or the license? In the past, you've had a pretty straightforward license. You pay $30 per seat, and there's only one product really to pick from from Oracle. You open up more choices. I'm just curious how you go about helping the customer sort through these choices. Thank you.
OK, great. The simple thing, we have a very well-defined playbook for our customers, partners, and Oracle that clarifies both. The first question we get, Cash, is, I'm an ex-customer in Y country. Which product from Oracle should I consider? Is it PeopleSoft? Is it E-Business Suite? Is it Fusion? By 11 or 12 different verticals, what we call a micro vertical, like a water company, and the top 40 business processes, order management, accounting, we have an explicit recommendation for them. That's completely transparent for them. Secondly, from a licensing point of view, the general question on on-premise versus SaaS, there are three choices that customers have. I can run the software on-premise. I can run the software in Oracle on-demand, which is a perpetual license that you buy, but it's operated by Oracle. Third, I can buy a subscription.
By subscription, I mean you buy a three-year contract, and then you just get a subscription-based offering, and it lasts for, and you can use it for the period of the subscription. There are two factors there. Generally, the things that we help them with is, are they comfortable that they have the skill set and the desire to run the IT function? If they say, no, I don't want it. If they say, I want to do that, then the answer is pretty simple. It's on-premise. If they say, I don't want to do that, then they have two choices. They can go with on-demand or they can go subscription. Do you follow me? Once it comes to that, then it's primarily a question of what their comfort level is in going with a subscription-based licensing, and that's largely based on capital versus operating expense considerations.
We have pricing calculators and other things that help them very quickly get to which path makes the most sense for them. Again, thank you all very much. I will hand it over back to Ken.
Thank you very much for being back. What we're doing now is we're going to switch to the latter part of the day. As I mentioned earlier, we're going to talk about some very important parts of our business, the support organization, which deals with a number of services that we provide our customer, whether it's software support or hardware support, ACS, education. There's a lot of stuff Chuck's going to obviously go over through that with you. A very important part of our business. From there, we'll segue to Edward Screven, who is our Chief Corporate Architect. We'll spend a little bit of time talking about Oracle Unbreakable Linux and Oracle VM. Then we'll move to John Fowler, who will then in turn talk about our hardware systems and the engineered systems that we develop. With that, let me turn it over to Chuck now.
Good afternoon. I'm going to go through a relatively brief presentation to do an update on primarily our support business. I'll talk about what's changed and how we deliver the service and how it fits in with Oracle's overall strategy. I'll also talk a little bit about the business side of it, kind of our revenue and quality metrics. Overall, I will leave this up for you guys to read. I know you're all fast readers, so I assume we're done. OK. Overall, support fits into a lifecycle of services that we have in the company that I assume you're mostly familiar with. The mission overall of our services for both software, hardware, and now for our systems is really to be able to protect the investment and to let the customers get the most benefit out of the investment they're making on the product side.
More and more, as you'll see through my presentation, more and more of what we're doing on the service side is linked into the products, both at the early development stages and also how they're deployed. This is, again, a simple listing of the different levels of service we have for our customers, starting with more of a traditional consulting service where we engage with customers to help design applications with them, solve specific problems that they have in their environments, leading down to support, which I'll talk about primarily. In my organization, we have a thing called advanced customer services. As you'll see, the support model is very simple.
We have one level of support, and we do have these advanced customer services, sometimes known as ACS, which will give customers more of a hand in providing support internally for their products or giving them more expertise as far as how they manage and how they deploy their systems. Also, as part of the support organization is Oracle University, where we continue to grow not only in the number of courses and the number of people educated on Oracle's products, but it's also a strong revenue and profit center for the company. Obviously, Oracle on Demand and some of the cloud announcements that you heard this week, where that's obviously for us a growing business, where we will host customers' applications, or in the case of products that are SaaS enabled, we'll run those products again as a service for our customers.
This is kind of the complete list of services we offer as a company. I'll briefly talk about primarily the support model. Oracle Premier Support is the one basic level that we offer for both software products, hardware products, and our systems. There are a lot of customers, and certainly on the hardware side, where companies have multiple levels. I'm sure you're all familiar with Platinum and Gold and Silver and all the variants of that with all the different pricing models. We have a very simple model at Oracle, which is this one level of support. It's called Premier. The Ts and Cs are available on our website. Customers or prospective customers can look at exactly what they are, what our response time commitments are, etc. It's all built into this one model of support.
That holds not only for the software products, but now as we've brought the Sun products into Oracle, we've moved our hardware to that model as well. When you look at our engineered systems, Exadata, Exalogic, Spark Superclusters, etc., they're based on this one level of Premier Support as well. I'm going to talk a little bit about these five areas as to what we're doing as far as changing how we deliver our service, how we deliver support to our customers, and why we believe it's an advantage for us not only in ensuring that our customers actually purchase support contracts and maintain them with us, but also more and more we're viewing support as a differentiator for our products. I'll talk about each of these five categories briefly.
First is that as we design software and hardware products, and this isn't specifically new, there are new capabilities that we keep adding with each release of the products. We design into the products the ability to diagnose problems, the ability to generate more information out of the systems about how they're running so that we can not only remediate problems as they arise, but we can also be proactive, get ahead of issues that could arise, notify customers about what modifications they should make, and then obviously avoid problems. The second major area I'll spend just a minute or so on is the fact that as we look at our engineered systems, we're providing what more colloquially is known as one throat to choke.
Obviously, as customers buy something like an Exadata system from all of the hardware components all the way up through a database or Exalogic up through the middleware layers and applications, we provide that one place where a customer can come and we triage the problems, we resolve the problems. It's obviously a much faster time to resolution because we have all the resources in-house. We have all the same processes. We share the same systems for the entire stack. It's a single point of focus for the customer, which many of them view, again, as a big differentiator versus buying components and having to run around and get support from multiple vendors. We also have and continue to enhance a number of proactive support tools. I'll mention a couple of those in a minute.
Also, we have for a number of years now continued to win awards, and it provides another point of reference for our customers, not only other customers that have been successful, but independent agencies, as well as analyst firms. If you look at Gartner, IDC, other reports, it's another point of comfort for our customers to see what their view is of our customer service and support. Also, continued innovation in each of these areas. Let's start out and drill down just a bit on hardware and software designed with support in mind. Part of what we've done is we've built things like, and you see on the list here, fault tolerance through how we do redundancy, how we do replication. This is, again, not only in hardware, but software as well. Our systems are engineered such that we can avoid these problems.
We can offer much higher levels of availability to our customers. We can also, again, respond much more quickly because we have certain things built in. For example, diagnostics. We do two things basically, which is, first of all, we know our customers' configurations. We have built-in agents sitting on our various software and hardware products that feed back to Oracle so we know what customers are running. If there's an issue with a system, the first encounter with Oracle isn't a long description of what they're running, what patch levels they have, what software is installed, what hardware rev levels they have. We really know what they're running based on the fact that they're connected to us. We, on a regular basis, in most of the cases nightly, collect those configurations.
The second thing is we now have systems built with the capability of alerting Oracle if a problem occurs or if a problem is about to occur. Again, more and more we're building into our products things that really help with the support experience. A single point of accountability or a single point of contact. We've done, I believe, a very good job, particularly with the number of acquisitions we've done, bringing all those support resources together using the same tools, using the same processes. If a customer happens to have Hyperion or VEA or a number of products that came to Oracle through acquisitions, they have, again, the same visibility as to what's going on. We have the same process as to how we communicate with them. We can really look at problems across the entire stack.
Our major tool for support with the bulk of our customers is a tool called My Oracle Support. It's not only the website that they come to, but it's a place where customers can file problems. In our parlance, it's service requests. If they run into a problem, they can file a service request online. As I'll talk about in a bit, with some of our hardware products, John Fowler and his guys have done a good job building in the capability for service requests to be filed automatically based on some condition in the hardware. Either the customer goes in, files a service request through a series of kind of guided steps, or the systems themselves automatically file a service request and get our support process going.
The other class of tools that are available on My Oracle Support are more in the preventive area, which is health checks, security checks, where, again, we, through collecting configurations that our customers have, we can look at what they're running. We can compare that to the database we have of our tens of thousands, hundreds of thousands of customers. We can pick out customers who are running similar configurations and what works best for them. If we think a patch needs to be applied to a system, if we think we can give them performance tips, they come to My Oracle Support. That is automatically available to them in a series of reports that are generated, again, based on their configurations. Some of the awards, again, we've talked about and the feedback we've gotten through some of these processes.
This is just one particular award from an organization that focuses solely on support organizations across the industry. It is what technologies are we using, the quality of our talent, the quality of our engineers, some of the experience that we have with our talent and our processes, and also the fact that we can do a holistic level of support across the whole stack. What you've seen this week at Oracle World is a whole other batch of products that we're delivering both on the software side and the hardware side. That has created, in my mind, some interesting opportunities for us to actually do a better job with support and help differentiate.
For example, if you look at the announcement John made a few weeks ago that was reinforced this week on the Spark Superclusters, you'll see a lot of what I'm talking about here embedded in those products and in that support process. Also, as we enhance and do new versions of our Exa products for analytics and certainly with new releases of Exadata and Exalogic, you'll see more and more of these kinds of features built into those products. I'll give one brief example. I'm not going to go through all the details here. Across hardware and software with the ExaSystems, in this case, Exadata, we have the ability for the customer to do a series of things that would normally be very difficult across systems that they just build from components from a variety of vendors.
For example, when we do installs, when we do configurations of systems, it's an easy task for us, an easy task for customers because these are standard systems. We've tested at the factory. We've tested back in development exactly the most efficient ways to configure these systems. That's what happens when they go out to the customer site. Also, when a problem arises, we know based on patch levels that are out because we do bundle patching for the entire system. You're not patching one component at a time. You're actually patching the system as a whole. We understand the state of the system and what customers need to do if they run into problems. Through these engineered systems, it's a much higher level of support experience for our customers. I'll spend just a couple of minutes more talking about the support business.
I know we've presented this slide, I don't know, probably 10 or 15 times to you. The idea is the support business model remains a very consistent, very strong model for us, as you can see in our financials every quarter. The model is very simple, which is customers buy a product from us, and we certainly collect the revenue for that product purchase. That starts this revenue stream over a period of time where we're building what in effect is a kind of an annuity where that revenue is repeated year after year after year. Most of our contracts are one-year contracts, and the renewal rate is extremely high. As we, again, build more product out, sell more product, that just builds that contract base. We've also made the point, I think, several times that independent of new product revenue, that revenue stream continues to come in for Oracle.
It's just enhanced by the fact that we're selling more product and adding to that contract base, which adds to the support revenue stream. If you look at software renewals, software renewals continue to be quarter after quarter, year after year, well above 90% in terms of renewal rates. The software model is such that obviously what customers get is they get new enhancements to the product. If somebody has bought a database and continues to make their support maintenance payments, then new releases, patches, enhancements, new versions are all part of what they get for that service charge. It's certainly in their best interest to stay current. The other thing, obviously, they get is they get the support experience and all the support resources addressing any problems they have. More and more giving them proactive advice about how to run their systems.
On the software side, the renewal rates remain extremely strong. I'll spend just a minute or so on Sun integration and what we've done, again, over the last several quarters as far as integrating Sun into this overall support model. The bulk of our customers now certainly on this one level of Premier Support that I talked about, and the attach rates continue to improve. Competitively versus where other people are in the industry, we feel very comfortable about the attach rates. We've spent a lot of time focusing on this in our sales organizations and in my support renewal organization, looking at kind of customer by customer what their situations are and moving them on to our standard Premier Support contracts.
The other thing is we're working very closely with our customers as far as renewing hardware and renewing and refreshing their environments, which then leads to continued support for us. As I mentioned earlier, on the software side, there's a strong component of the support offering that they get, which is new versions, new updates to the software products. On the hardware side, as people cycle out of hardware, we have a major effort to make sure that what they cycle into for older systems is back into our systems, whether it's Sun Microsystems systems similar to what they were running, or also more and more now into the Exadata systems. That's a major effort and a major contributor to the hardware support revenue stream.
We've also spent a lot of time over the last 12 months making sure that the experience that customers have is very strong and is as integrated as possible into the general Oracle support experience. When customers come in with a problem, the support experience they'll get for hardware as well as software will be at the same level, and we will triage issues for them across that. The Sun Microsystems customers now have an integrated experience. They know they're dealing with Oracle. They know they're dealing with processes that they were very familiar with on the software side. That was, again, one place to come to solve all their problems and to get advice on how to make sure their systems stay up, available, reliable. We also have done a lot of insourcing of resources. That was a major effort that we made with the acquisition of Sun Microsystems.
Sun Microsystems had outsourced multiple levels of their support organization to third parties. Our insourcing has had two major benefits. One of them is from a productivity and a cost standpoint. We've dramatically improved our profit margins on hardware support based on the fact that us to do this versus what they were doing with third-party outsourcers is significantly cheaper. The second thing is the experience for customers is better because they're getting Oracle employees, people who understand Oracle, who understand the overall processes much better than third parties did. We view that as a major success as part of the Sun Microsystems integration. Overall, it's going very well. You can see again in our financial statements the result in revenue and margin certainly, and that behind that attach rates are moving at a level where we feel very good about the progress.
The model is Premier Support, one level of support. In the case of the hardware systems, 12% annual fee gets you support for the OS and all the hardware. 8% gets you support for the software, which is the operating system and then Oracle VM as well. For both of these levels of service, very good improvement in the attach rates. The last slide I'd just like to leave you with is all this leads back to one metric for us, which is the offerings that we make to our customers, the level of support we provide. The feedback to us is not only did they continue to renew, to buy our products, to renew their service, but also we do very detailed customer satisfaction tracking.
Quarter by quarter, year by year, and this includes the acquisition of Sun and a whole set of new customers and new transactions on our hardware products, we continue to be above 90% customer satisfaction in all periods. That's a quick summary of where we are with support. We'll have a few minutes for questions later. With that, I'd like to introduce Edward Screven.
All right. How many times does this appear today? OK. I'm kind of wondering, like, you know, at what point does it get so there's so much text on here that it just turns into DOS. I'm going to talk to you today about three things: Oracle Linux, Oracle VM, and MySQL update. Oracle Linux is Oracle's comprehensive Linux offering. We offer enterprise-class support. We have training. We have consulting services. It's the only Linux a customer needs. We engineer it together with Oracle hardware and software, but it's designed to run on lots of vendors' x86 hardware. We have certifications on other vendors' hardware. It's designed to run any Linux workload. It features the Unbreakable Enterprise Kernel, which is fast, modern, reliable, and optimized for Oracle. It's compatible with Red Hat Enterprise Linux. Our customers can switch to Oracle Linux without reinstalling.
You just take your existing Red Hat servers, you point to our update stream instead, and you're done. It is Oracle's Linux choice. It's what we run. It's the Linux that we run in Exadata and Exalogic and Exalytics for extreme performance. It's what we use inside Oracle, exclusively inside our product development organization. We have more than 20,000 engineers who develop our products on Linux, on Oracle Linux. We do not have Red Hat Enterprise Linux inside our test farms or development farms. We do not test our products on Red Hat Linux. We certify on Red Hat Linux to date because we know that Red Hat Linux is compatible with Oracle Linux. It's the only Linux that we recommend for use with our software. The Unbreakable Enterprise Kernel is something that we launched last year. It combines the best of community enhancements with Oracle innovations.
We designed it to be scalable. We designed it to be reliable. We added additional enhancements to do things like very low latency, very fast, and thin-of-fan networking. When we benchmarked it against our compatibility kernel, it was 75% faster. This year at Open World, we launched a second release of the Unbreakable Enterprise Kernel, even faster at scale. Important improvements like a better scheduler for high-thread count applications like Java, transmit packet queuing for low latency networking of all kinds, transparent huge pages for large memory applications, something called containers. Containers is a feature that Solaris has, which is very useful for doing low overhead isolation of applications, a virtual switch that lets you do things like filtering, quality of service, VLAN isolation, and DTrace.
Again, something from Solaris that helps customers understand all of the events that are happening inside a running Linux kernel, Unbreakable Enterprise Kernel, that may affect their application. Now, a Linux system admin faces a daunting problem on a regular basis, which is that a security errata or other critical update comes out that has to be applied. That means that I need to take down the applications. The first thing they have to do is negotiate with a line of business owners to pick a date, days away, weeks away, in which they will be allowed to take down the application. When that date finally comes, they have to bring down all the application servers, bring down all the databases, apply the updates, bring the application servers back up, bring the database back up, do sanity checking.
Finally, days later, weeks later, with significant application downtime, they finally applied that critical fix. During that entire time, they were vulnerable, vulnerable to a security issue, vulnerable to a data corruption issue. That problem is solved now with Oracle Linux. We have a technology called CaseWise that lets us take an update to the kernel, like a security errata, turn it into a special form that we can apply online to a running Linux kernel without taking it down. No rebooting the machine, no taking down those application servers, no taking down the databases. Applications running on top of Oracle Linux, to which we apply one of these patches, have about a millisecond pause. There is no disruption to the running application. I can tell you, we've briefed dozens and dozens of customers, me personally. This is extremely appealing to them. This is a unique feature to Oracle Linux.
Only Oracle Linux offers zero downtime patching for critical updates. It's available today as part of our Premier Support offering for Oracle Linux. If you compare Oracle Linux with Red Hat Linux, I can tell you objectively, Oracle Linux is by far the best offering. It's the best technically with the Unbreakable Enterprise Kernel, with CaseWise. It is the best in terms of support. We have excellent enterprise-class support that Red Hat does not match. It is the best in terms of price. That's why a customer like Dell has Oracle Linux as their default Linux platform. Linux has a long history at Dell. It started with Red Hat back when it was called Advanced Server, I think. They switched to SUSE, then they switched to Oracle Linux. Today, they run 5,200 servers on Oracle Linux. When they do a new application deployment, it's on Oracle Linux.
They're not the only customer that selected Oracle Linux. We now have more than 8,000 customers. Now let's turn to Oracle VM. Oracle VM is Oracle's high-performance server virtualization. We have state-of-the-art virtualization features like live migration, high availability, template deployment, automatic power management. It lets you virtualize your entire data center. You can run guests that are Oracle Linux, Windows, Solaris. It runs on both x86 and SPARC. It's engineered for Oracle, again, so it's engineered together with the hardware and with the software for maximum performance and reliability. It's still built out as a horizontal offering. Customers can run it on other vendors' x86 hardware. They can run other workloads other than Oracle workloads. It's certified for use with all Oracle products. We just released in August a significant new release of Oracle VM called Oracle VM 3.0.
We have significantly enhanced Oracle VM Manager, so a much better user interface, policy-driven management, which means the system will automatically help you save power, automatically optimize the use of resources. You can centrally manage both networking and storage associated with Oracle VM, which is actually a very powerful feature. You can manage thousands, thousands of VMs, your entire data center. We are very focused on performance at scale. An Oracle VM guest can have up to 128 virtual CPUs, up to 1 TB of memory. I don't think today we have someone presenting Enterprise Manager 12C. Enterprise Manager 12C includes a very interesting new functionality called Cloud Management.
It lets customers automate the entire lifecycle of a virtual pool, helping customers do things like creating templates based on Virtual Assembly Builder, helping them do chargebacks because a lot of customers want to do chargebacks to lines of business depending on what's running in the pool, helping them do self-service provisioning. If someone from a help desk doesn't have to go and do something to the pool to bring up an application on behalf of that line of business. That is now integrated with Oracle VM 3.0. With that, it becomes very easy for Oracle customers to create these private clouds. The fundamental difference between Oracle VM and VMware is that Oracle VM is built to be part of a complete stack to make applications fast, reliable, and easy to deploy and manage. VMware is a point solution.
It's a very functional point solution, yet it's still a point solution. When you manage through VMware, you are not managing the operating system, you are not managing the hardware, you are not managing the middleware, the database, or the applications. When you manage Oracle VM through Enterprise Manager, you are managing all of those things together. The middleware, the database, the applications from Oracle understand virtualization through Oracle VM. When we build Oracle VM, we specifically build it to run production mission-critical workloads of all of that software stack. Oracle VM is four times more scalable than VMware. VMware is four times more expensive than Oracle VM. In fact, that's actually a conservative number. Here's a pretty typical data center pod that our customers might have: 250 two-socket servers, 96 GB of memory apiece, pretty common.
The three-year cost for Oracle VM is $500,000. The three-year cost of VMware is $3 million, more than six times more expensive. Hasan, I think, discussed quite a bit about Oracle VM being embedded in Exalogic. Of course, it's also embedded in our SPARC Supercluster. Oracle VM for SPARC is embedded in SPARC Supercluster. It enables customers to run mixed workloads on a Supercluster, even different versions of Solaris on the same box. One of the advantages of having both Oracle VM and the hardware in-house is we can help engineer those things together. On T4, Oracle VMs for SPARC can take advantage of facilities in the T4 for dynamically assigning physical CPU resources to threads and help our customers optimize and choose between single-thread performance and total system throughput. We're able to take advantage of other T4 facilities like on-chip encryption.
When we live migrate a VM from one node to another, we do that encrypted using hardware encryption built into T4. Finally, MySQL. We have a really straightforward MySQL strategy. We're going to make MySQL a better MySQL. I pronounced it both ways in the same sentence. I think that's kind of astounding. OK. All right. What I mean by that is we're going to improve it. We're going to make it better without losing the attributes which actually made it very popular to begin with, which is it's small. I mean, it's not very large in terms of the bytes it occupies on disk. It's easy to install, and it's very easy to administer. We're going to enhance. When I say make it better, we're going to enhance both the Community Edition and the Enterprise Edition. We're going to make it better on Windows.
That means dramatically improved performance on Windows and dramatically improved the integration with Windows, making it a more natural Windows product. We're going to make it part of the Oracle stack. A lot of customers, many customers, probably every one of our large customers use both Oracle Database and MySQL. They have a lot of surrounding infrastructure like Enterprise Manager or Secure Backup, Audit Vault. We're building and have built integrations into those products to make MySQL a natural part of the Oracle ecosystem. Since we acquired Sun Microsystems, we've had a steady release of new MySQL releases, both the core server and the components that surround it. If you're kind of a MySQL cognizant person, you would know that these releases have been at a quality level far beyond what was ever released under Sun Microsystems or under MySQL itself.
We're applying disciplined engineering techniques to MySQL to make MySQL better and also improve the experience of MySQL customers. I mentioned before that we would enhance both the Community Edition and Enterprise Edition. Just to make sure folks understand what that means, Community Edition is the version of the MySQL server that we released under the GPL, and anyone can take and use for free. MySQL Enterprise Edition is the version of MySQL that people pay for. MySQL Enterprise Edition gives you the right to use tools like MySQL Workbench or Query Analyzer. It gives you the right to get support, of course. Now, it also gives you the right to use features in the core MySQL server that do not exist in the Community Edition. One of those features, just as an example, is something called thread pooling.
Thread pooling allows us to more efficiently assign compute resources in a MySQL server when there are a large number of clients. On this chart, you can see the OLTP read/write performance of MySQL with and without thread pool. With thread pool, as the number of database connections goes up, you eventually kind of plateau and stay pretty much flat. Without thread pooling, your performance drops off precipitously. Now, for the typical community user, the person who's not paying and frankly is never going to pay, that doesn't matter because the workloads they're running don't use large numbers of connections. For our commercial customers, that scalability matters a lot. Most of the transactions that we do in MySQL, the commercial transactions, they tend to be small. I mean, a small number of units embedded in a relatively small department of deployment.
That is not the only kind of deal that happens in MySQL. All of these customers have bought MySQL in the last year. The average deal size here is close to $1.5 million. All of these customers also use Oracle software. Some of these transactions happen in conjunction with Oracle deals. For example, Huawei bought a whole lot of Oracle Database and a whole lot of MySQL all at once. The reason why they did that is that MySQL and the Oracle Database are good at different things. By adding MySQL to our portfolio, we expanded the database market we can address. That is the end of my presentation. Thank you. Actually, I got to do one more thing. Sorry about that. I have to introduce John Fowler, who runs our Hardware and Systems division.
OK. Obviously, we rehearse repeatedly before these. I'm going to talk about hardware and give you some of the background and how we're thinking about hardware and what we're doing in both the strategy and the products real quick. This will not be a microprocessor design course, although you might feel that way after you read this particular statement and any comments about future product releases. A little bit of background here in terms of how we think. One of the things that's important to understand is we think about being one engineering team. Obviously, we are tens of thousands of engineers and billions of dollars in R&D. Here at Oracle, we have microprocessor development, systems development, storage development, operating system, virtualization, database, middleware, ERP, CRM, industry verticals, development platform, development tools, management tools, the whole nine yards.
The way we think is how do we build best-of-breed products in each of those technology areas. Larry is unwavering on that. You can have a best-of-breed core technology component. He's also unwavering at how do we build things together better so they operate better for better value for customers. The scope of this now changes to include hardware. I'm particularly excited because we've actually been part of Oracle long enough now to start to really feel our legs underneath us in the products that we're developing along this principle. Engineering does take a while. This Oracle Open World is kind of a coming-out party for the products that we've now built with enough time to have the influence of one engineering team. It's tremendously exciting.
One of the things we talk about is a very simple idea: how do we take this one engineering team and produce in core products things that have 10 times the performance and 10 times the value of products you can get in other ways using à la carte computing. Value comes in lots of ways. Performance is value as well because you can use less. It also can be reliability. Is it easier to manage? Can you set it up quicker? Do you have better service and support and so on? We're cohesively taking these different engineering elements and putting them together to create step functions in performance and capability as opposed to just the incremental leaps in capability that you've been used to for many, many years now in the hardware industry. The advantage here is a very simple concept.
For several decades now, people have been assembling their infrastructure out of selections of different components from different vendors, storage vendors, and even virtual.
Point products and everything else. That has actually worked sort of remarkably effectively for a long time. What we're doing is taking the fact we actually produce and engineer all these core components together to produce better solutions. We're not thinking about these as bundles, so to speak, which is I have the same stuff, so let me put it in a pile and ship you the same thing. We're actually looking at changing different parts of these technologies, not just testing together. For example, I test the Oracle Database for millions of hours now on my processors before I even release the processors to the market. That's kind of an obvious thing we can now do that was never able to do before because the Oracle Database and the processor design team were not in the same group. Also, can we change the relationship of database to storage?
Can we build middleware differently and so on to actually produce a different technical result than you can with the separate parts? Certainly focused on that. The pinnacle of this effort is engineered systems and appliances. We have been spending a lot of time in the marketplace talking about Exadata and the other engineered systems. I just want to be clear, we continue to build best-of-breed à la carte products. Customers want to buy a server, which is what they're buying today. They can buy a server. We continue to increment the component technologies as best-of-breed technologies and have huge enhancements that we've announced over the last couple of weeks. The pinnacle of where we bring together all this co-engineering is in engineered systems and appliances. We came out with Exadata and have established that for some time now, purpose-built for the database, Exalogic, purpose-built for middleware.
At this particular show, we talked about big data, where we bring together key value pair, NoSQL data with Hadoop to allow you to look at data in a different way, Exalytics, which accelerates business intelligence, and then the Spark Supercluster, which expands the scope to general-purpose computing. The focus here is bring the engineered system concept of acceleration, but expand the scope. If you look at this and understand that what I can do now is I can run dedicated OLTP, I can do data warehouse, I can do business intelligence, I can do middleware, I can do packaged applications, ERP, CRM, I can consolidate legacy applications, as well as doing big data. Essentially, we can now apply Oracle's engineered systems to the entirety of the core problems that face most enterprises today.
We have moved way beyond starting out with a data warehouse, expanding data warehouse to OLTP, to now covering essentially the whole enterprise. The 10 and 10 rule applies quite dramatically to Exadata and Exalogic. Remember, Exadata is purposely built for a database. What we did there, we engineered into the storage itself some database logic. This in one machine brings together network storage and servers, as well as database 11. We can get 10 times the query performance of a conventional storage architecture, 50 times the warehouse performance, and so on. You do not need other storage. You do not need other networking. You do not need other servers. The other part of the strategy, obviously, is by having a great deal of integration and then Exalogic around the cloud. One of the things that we do spend time talking with customers about is how to use these together.
We have done specific engineering, where if you lay down an Exadata and then you add an Exalogic, we have made those also work together better. You get additive properties of performance and extreme scale. Exalytics was just announced this week. Very similar idea here. For customers using OBIEE, the BI Foundation Suite, Essbase, and the tools on top of technologies like that, like Hyperion, get a dramatic in-memory acceleration for business intelligence. You can use this as a standalone product. For example, you do not have to be using the other parts of our family. If you do use it with the other parts of our family, we actually connect up to the same high-performance fabric. We use the same management tools. It becomes an integrated part of the whole. You get extraordinary improvements in business intelligence performance. It is pretty clear.
Larry, in his keynote, talked about this notion called the speed of thought. I frankly do not really know exactly how fast the speed of thought is. What I can tell you is this is really about moving BI from a reporting concept to a real-time concept. This is a huge step forward in in-memory computing for business intelligence. Spark Supercluster was announced about a week ago. This is about bringing together for our Spark Solaris customers a consolidation platform that does take advantage of a number of these core technologies to then create a brand new stepping stone for running traditional applications, whether packaged or legacy, together with the Oracle database. What we've done is we've taken a new processor called the T4, which is the biggest single-generation boost in performance that we've had in our history.
We've combined it with the core IP of Exadata, which is the accelerated storage cells, with Solaris 11. We allow you to run the core IP of Exalogic to run the acceleration of middleware, as well as a number of other enhancements to give you a single platform which is virtualized to go run an extraordinarily accelerated database 11 alongside an application such as SAP, PeopleSoft, SaaS, applications that you build yourself, eBusiness Suite, and Siebel. At the show here today, we showed, for example, Siebel, PeopleSoft, and eBusiness Suite together with Oracle 11 serving 30,000 users on a single system, Supercluster. 30,000 concurrent users on a single system, both batch and concurrent simultaneously, which is quite significant. This is actually an example of this particular environment. One of the things to understand is the distributed parallel nature of the hardware.
What we've done with the Oracle database and the storage means that we're no longer solving problems through just brute force. We actually parallelize all of the SQL queries. We've integrated high-speed networking and flash in a unique way to give the kinds of I/O performance and throughput performance that, frankly, it's impossible for a conventional server and storage infrastructure to equal, except at fairly ridiculous costs. The trick here, and this is an example I'm not going to spend too much time on, but it's important to understand. When you look at Exadata, you say, why is it so fast? We did not solve this through brute force. We solved this because we have a combined OS, database, and hardware engineering team.
We designed a different kind of storage unit that effectively puts a portion of database logic into the storage unit so that when you execute a database query, the database query is actually farmed onto the storage units. The results are aggregated back. This is why I say there's not a brute force approach because it's not dependent on growing the biggest possible server technology or the biggest possible storage technology. It's because we've applied a degree of intelligence to this. Like our other products, though, you don't have to change your applications. It's very important. Our customers are not changing any applications to take advantage of engineered systems. Similarly, for the Elastic Cloud, we've done work in the networking layer with our network in OS and in Java and middleware to provide similar kinds of performance increases for Java and JMS.
What this adds up to in Supercluster is, and again, this is somewhat an unfair comparison. I say it's unfair because we're not using brute force anymore to solve problems. If you wanted to equal the $1.2 million I/O operations or the 2 million JMS messages per second or any of the other things I showed you in the previous slide, you would have to build a system so big from IBM that you would need to borrow money from Larry to afford it. It's just, again, a difference in our approach. This becomes a sustainable IP advantage because as we increment our components, it scales with us. As we make storage faster, as we make the processor faster, as we make the networking faster, the overall capability of the system will continue to go up because of the way we've done it.
There are a lot of people using these systems now. The thing I want to emphasize now is our scope has changed. We started with data warehouse. We expanded to OLTP. We started to go into middleware. What we are now poised to do is to cover every important technical problem in the enterprise. We already have examples from CRM and ERP and middleware and others, people in deployment using these technologies. It's an important thing to understand the basic vision of engineered systems. These are products that you can purchase individually, and our customers can use with all their existing apps. You can use Exalytics with all your existing constructs on OBIEE and Essbase. You can use Exalogic with all your current Java and WebLogic J2EE applications. You can use Exadata with every single application you use in Database 11. You can also use these products together.
What we're seeing customers do now is they started with Exadata, they've had Exalogic, and now they're looking to add other members of the family because we've engineered them to also work together. One of the core technologies that we introduced in hardware today goes back to what Larry has been very focused and gives me a lot of attention on, is making sure that we have best-of-breed technology components as individual products. We want to win at both levels of the system. We introduced and are now shipping the SPARC T4 servers. It's the biggest single-generation 5x boost that we've had in our history. It really catches us up and puts us on top of Intel and IBM in the core server place. It puts us at a really interesting price point for very powerful systems that have virtualization and operating system built in.
I'm going to come back to that in a future point. There are a number of interesting things about these systems, such as that we have some core technologies specifically around technology called dynamic threading. We are really the first ones to construct a processor that can handle what's called a very wide-range throughput workload, which is commonly things like the database, and then a very strong single-thread workload where you want low latency. What customers have had to do historically is pick different servers and different technologies to build for different purposes. What the T4 lets us do is effectively give us a universal server. That's what makes a Supercluster a universal consolidation platform for different performance levels. I want to shift gears and talk about Solaris 11. Obviously, one of our key value attributes is we can innovate at the operating system level as well.
We've been previewing Solaris 11 with customers now for some time. This is an effort that has spanned years, thousands of projects, hundreds of patents, and is a very significant enhancement to our overall Solaris effort. First of all, one of the things that's very different about us from other Unix vendors is Solaris runs on both Intel and it runs on SPARC. That allows customers to go and have particular affinity to the Unix platform, choose and use servers from different technology areas, different performance characteristics, and have great freedom in terms of how they actually implement their infrastructure. We do the best possible effort on both these platforms. With SPARC, we design the hardware as well, so we have an opportunity to put features into Solaris that we also put into the silicon. Nonetheless, we work hard on making it work on Intel as well.
I want to give you the next 40 minutes we'll spend in a detailed technical treatise on latency-aware kernel memory allocators. The point I want to make here is we've been part of Oracle now for a while. One of the extreme focuses here was to go put significant enhancements in Solaris to run the Oracle stack. These are just a few of the kinds of technical enhancements. When we work in concert with a database team, we really understand how the database works. We are free to go and make changes directly into Solaris and test them to go and make the database run well. Of course, a number of these are center points to how we work well in a Supercluster environment.
Shifting gears to storage, again, the storage idea is very similar to the others, which is build a best-of-breed technology components that you can purchase separately, but do extra integration with Oracle. We actually think about storage a little bit differently than other storage vendors because we have some different pieces. The first thing is we do think about a concept called database I/O. This is what's encapsulated now in Exadata, as well as Supercluster, which is really a storage system of our own design that incorporates database logic and offloads query processing from the conventional storage server. There's simply no way in a conventional storage system to equal the database I/O performance that you can do when you actually parallelize SQL queries and actually execute them in Exadata storage cells. For us, that's a category.
That's obviously the most premier way to run the database because you run so much faster. We have products in NAS. We recently acquired Pillar Data Systems for the SAN environment. In both of these products, we work on ensuring that we have key features that make them highly competitive or superior in their environment. We also incorporate elements of core Oracle IP to create an advantage when using Oracle software. As an example, we announced last week that we've incorporated hybrid columnar compression for the Oracle database into these storage units. What this means is any customer of the Oracle database that then hosts their data on either Pillar Axiom or on the ZFS storage appliance can achieve significant cost savings because the database can automatically be compressed by the inclusion of this feature capability.
Obviously, we have StorageTek, which is the world's leading enterprise tape automation technology with enormous scale and reliability and has been the leader for literally more than a decade now. If we look at the Pillar Axiom 600 storage system, the classical competition is against EMC. Against EMC, we have a more scalable system because it's built on a different architecture, which allows you to scale up to eight controllers instead of two. We offer much more effective utilization efficiency because we can go up to 80% utilization with the way that the architecture is between RAID and the controllers. We've incorporated the hybrid columnar compression to be more efficient and effective with the Oracle Database.
One of the unique features of Pillar Axiom that EMC and, in fact, no other mid-range SAN vendor has is in a shared environment where you're going on a storage network and you're being used by many hosts, we have an outstanding quality of service capability, which allows you to decide what SLA you want to have with which host, which is perfect for, for example, deploying in a SAN-based cloud environment or any kind of environment where you want shared tenancy, for example, between transaction processing and other environments. The ZFS Storage Appliance, our NAS offering, logically lines up with NetApp. We've been doing some very significant enhancements here as well. On a cost performance or an absolute performance basis, it's far superior to NetApp. Obviously, that's not the only choice that people make when they look at storage.
We also have superior understanding, especially in a cloud environment, through our analytics on what's actually happening with different client environments. Like Pillar Axiom, we've incorporated compression, hybrid columnar compression. The ZFS Storage Appliance has the additional capability of connecting directly to the InfiniBand fabric that is being used by Exadata, Exalogic, SuperCluster, and the other tools. You have an extraordinary performance storage directly connected to the engineered systems using the InfiniBand fabric, which is five to eight times faster than any other storage fabric in the world today. Hardware management is extremely important to our customers. Our customers, by definition, do tens, hundreds, or thousands of systems. What Sun had was a set of products originally called Opcenter, which allow you the base level of hardware infrastructure management, including virtualization.
Operating system, health monitoring, telemetry, patching of the operating systems, deployment of virtual machines, and then having a console to manage all of these capabilities. One of the things we recently announced is this core product is available for free for anyone with a system support contract. It is important to understand the ramifications of this. When you buy an Oracle server product, you get virtualization included. You get the operating system included. You get no limits in terms of the number of virtual guests you want to run. We do not have for the operating system and virtualization fancy per-core licenses or per-memory size licenses or per-phase-of-the-moon license. We do not have any of those license terms. You buy a server. You get virtualization. You run as many virtual contacts as you want.
We now include all the core management capability to go provision and manage all your virtualized infrastructure. People who are buying point products, for example, like VMware, have a really interesting new economic opportunity to go look at, which is just buy a server from me and get all of that stuff included without limit. We are pretty excited about this. The hardware management is not the complete management story. Obviously, we have the lifecycle management of the hardware. Our vision for managing both virtualized environments and clouds extends far beyond that. The first step is we understand and know the applications. We can manage database middleware, the packaged apps, and all the technology above. You do not have to use my hardware to take advantage of those.
If you are using my hardware and lifecycle management together with Enterprise Manager for the applications, then I can tell you, for example, through the stack, if you have problems and how you manage the lifecycle. Since then, we have also added, and Edward Screven talked about this, a real great understanding of business-driven application management, which is how do I self-provision? How do I know if there is a quality of service issue, for example, that Siebel uses? What is going on? We are unique in that we have knowledge of the app, the middleware, the database, the operating system, the virtual machine, the hardware, and the storage. Therefore, we can express that complete lifecycle to a customer and greatly simplify their environment for running their cloud-based applications. We're, of course, fond of talking about million-dollar machines. Those are useful things.
It's important to understand that this is all part of an architecture of software and hardware that reach down into and include departmental-level customers and small customers. What we do is make sure that if you're growing up on WebLogic, you're growing up on the Oracle Database Standard Edition, you're growing up on Solaris or Oracle Linux, you're actually growing up with really the exact same software that the biggest people run. You can just grow all the way up that stack and use a common set of tools, a common set of capabilities to reach the highest levels of transactions, business development, business intelligence, and other kinds of environments, as well as creating opportunities for partners, mid-market, channel, and departmental. Sun's been a part of Oracle now for a while.
One of the things is at this point, Oracle really runs a huge percentage of their business on Sun hardware. Larry has a very interesting philosophy about that. He makes me help run it. Every single day, my teams find out exactly what's good and bad about our hardware. Effectively, Oracle IT and development is integrated into development engineering. I have one of the biggest petri dishes in the world to go improve my operating systems and my storage and my physical hardware. That's not always fun because the folks in Oracle can be very blunt. I can tell you this has made my products dramatically better. Finishing up, best-of-breed compute and storage building blocks. We had some huge announcements this week around a whole variety of hardware products like Exalytics, T4, T4 Supercluster, the third generation of ZFS, Pillar Data, and so on.
We co-engineer these with the software stack and build engineered systems. These engineered systems now can span all of the core problems that an enterprise would care to face, from legacy applications, packaged applications, middleware, database, data warehouse, business intelligence, and so on. With that, I'd like to thank you for your time. We're all done.
Hi. Thank you very much, Kashwan. I'm in Maryland. Question for you, John. Throughout your presentation, an overarching theme seems to be optimizing the Spark architecture with the help of engineering the database simultaneously or vice versa. How does this play out in the broader market? Should we think about your strategy going forward as something along the lines of, OK, Spark's really going to be a great fit within the Oracle install base, broader market? How does it play out? I mean, that's something that's on my mind.
Yeah. I mean, we're using the Oracle Database and other technologies to drive Spark really hard. There's a learning loop that's going on with there. I've already talked publicly about some of the directions of Spark, which include actually embedding some software features into silicon that will go and help accelerate enterprise applications. These include embedding some basic data type support and embedding compression and so on. I've been, of course, very careful to say when these will show up and so on. What I will say is those will also benefit other applications. We're using Oracle software to help drive the development of those. That doesn't mean that some of those won't be highly beneficial to other enterprise applications as well. We're not focusing Spark really on high-performance computing or any kind of thing like that.
It's really what is the best possible enterprise processor we can make and how can we drive a feature set by working with the best enterprise software team on the planet to actually make that happen. The benefits aren't purely for Oracle software.
Hi. Kirk Materne with Evercore Partners. I guess, Edward, when you look at the virtual, I guess, server virtualization, at first, a lot of the movement was on low-end file and print servers and dev and test environments. Two questions around that. One, do you now see a lot more movement around high-end workloads, around databases and applications moving in that direction? How do you convince the people that built their virtual infrastructure to move over to a different set of tools for these higher-end systems? Do you see that as there's an inflection point where you can not only move them over to high-end systems using your tools, but also bring over the workloads that were on file and print servers originally?
First of all, I think you're right. There is definitely a shift in customer interest away from those kind of lower-end workloads or like dev and test instances to production systems and consequently also to larger systems, bigger databases, bigger application server instances. I think the way that we shift the install base away from VMware, and let's face it, that's really where it is right now. The way we shift away from VMware to Oracle VM is, number one, demonstrating superior scalability performance. Number two, integrating it into our engineered systems. When a customer buys Exalogic, when they buy a Supercluster, Oracle VM is there. That, I think, is also going to help give customers the confidence they need to run mission-critical applications on Oracle VM. Finally, there's just the value question. We have customers who are paying huge, huge VMware support bills.
I don't think there's one customer that I've talked to who isn't interested in trying to find a way to reduce that. Oracle VM certainly does that. It's a zero-license cost product. If you use it on hardware that's not Oracle, you pay a very modest support fee. If you use it on hardware that is Oracle, then you pay nothing for support as long as you have system support.
Hi. Hi. Over here, Dan Fletcher with Neuberger. John, if I can ask you about storage, how would you, on a scale of 1 to 10, rate your advanced storage features, deduplication, replication, and so forth? If it's 11, why haven't we seen more?
Yeah, 11. First of all, we have several different storage products, so it's a little hard to answer. Obviously, in Exadata, since we're able to both do row column compression and encryption and operate on it with OLTP-level transactional performance, that's more like a 25 on a 1 : 10 scale. Simply, nobody else can do that. It's simply impossible to do. We're doing local storage processing to all do that. In the case of, do you want me to walk through all the products? I mean, that's database I/O. In the case of NAS and SAN I/O, when it comes down to obviously working with the database, again, we're going to be essentially off scale since we have a unique feature that allows you to compress the data and then use it in a database environment subject to some performance penalty.
We're outstanding in replication, clone snapshot as compared to NetApp and EMC. Deduplication is available as well. It's a little hard to rate deduplication because it's a long topic. I think we do quite well there.
I'm going to add one thing. I think it's important to keep in mind that a really large portion of high-end storage is Oracle data storage. The advantages that we have for Oracle Database kind of translate instantly into storage advantages in general when we sell storage products.
Yeah. One other way we are going to confuse the market just because we think a little differently is in Solaris 11 with our own file system and other people's block storage, we provide deduplication as well as compression. You can use both of them. You do not actually have to get it from the SAN vendor if you do not want to. Is that a storage product or what? I do not know. We stuck it in the operating system, and it is available for every server running Solaris 11. I think we are in pretty good shape there.
Can you give some examples of other software vendors that are validating their software applications against your hardware? You mentioned SAP. How broadly is that happening, if at all?
Yeah. I presume you're talking about the new stuff like T4 and Supercluster. We have a broad effort going on. Obviously, we have all of our own stuff, which is a relatively large software portfolio, I've heard. One of the ones that's already done a public endorsement is SaaS, for example. SaaS is here on the floor. We're working with SAP and a pretty broad range of software vendors. We have a particular program that includes engineered systems, all of the engineered systems, not just Supercluster. It includes Exadata and Exalogic to go work with all the main software vendors to make sure that their application software are validated. What we find is this is really about producing sizing guides. Software works just fine. The sizing is particularly interesting when you're 10 or 20 times faster. How big does the database actually need to be?
How many clients can you support? It's an extensive effort. You'll see a lot of public statements from companies about the support of that.
Thanks, Rick Strillen. The question is on Spark. Just how strategic is it to Oracle? If you could sort of address how you compete against Intel longer term, are there advantages you have because you have the rest of the stack and you can do things in the chip that Intel's not capable of doing?
You got the answer already.
Do you do Trigate and 14 nanometer?
Yeah. Let me just be perfectly clear. First of all, we take advantage of Intel. We use Intel both to build systems and put our software on. We're constructing this as a win-win situation where we take advantage of Intel where it works well for our products and others. From a semiconductor standpoint, you got to understand I'm not all like Intel. I'm not selling chips. I'm not in the merchant silicon business. OK? I'm building systems. As a side effect of building systems, I can go design the system how I want. I can design the silicon how I want. I can design the software how I want. What you saw with SPARC is we went through some pretty turbulent waters at Sun Microsystems and fell behind in some of the R&D development. At Oracle, we're able to go and invest and work in a more focused way.
You see the first result in T4, which is this 5x improvement. What you said is exactly right. We're investing in core features. We then make the features work with our operating system, virtualization, database, and so on, whether it's cryptography, acceleration, or others. That's how we compete because the customer's not buying a chip. They're buying applications. They're buying application performance. One last comment on that, and people always get Intel and us mixed up, is I don't have fabs. I don't spend billions of dollars in fabs. People say, how are you going to keep up? I don't do any of that. All I have is a design team. It's kind of like having a software team. Our fabrication is at TSMC. What we found is TSMC is absolutely world-class in fabrication technology. They have to go spend the billions. I don't have to do that.
Let's not get that part mixed up.
Hello. It's John DiFucci from JPMorgan. Questions for Chuck . Chuck, you mentioned that attach rates for hardware support's been improving. I'm not sure that Oracle's ever talked about what those attach rates are today. It's my understanding anyway that Sun Microsystems had sort of a unique relationship with partners providing some kind of support along the way, and that's been changing. If you could talk a little bit about there's two opportunities here. There's the new sales and support on that. I would assume that's much higher than what's in the installed base. At the same time, that installed base sounds like it would seem that there's a lot of low-hanging fruit once you stopped what was happening before. If you can talk about perhaps how that's going and how much further you have to go for that.
Sure. There haven't really been attach rates that we've released specific numbers. The industry generally doesn't. I don't think anybody has actually given them. One of the reasons is it's relatively complex. You can't just give a number and understand how it's going. Let me explain to answer your question some of that complexity. You're absolutely right. When we sell a product, the assumption is that our salespeople who are selling those products, the product salespeople, will attach support to that. For the direct business where we're selling directly to the customer, that is now incredibly high. I think we've really done a good job with our internal sales force. They get it. Their commission's on it. That's working very well. A lot of the business at Sun, and less so now, but still a big portion of our business is sold through partners.
The partners, we're still kind of unwinding a number of agreements where partners don't necessarily attach software immediately. Some of them do it within 30 days, some within 45, 60. It depends on what part of the market they're in, for example. That's been a little bit slower, but getting the improvements there. It's a little bit behind what we've been able to do with the direct sales force. That's what we would consider point of sale attach. The next couple of categories are after point of sale attach, where we go back to a customer that didn't buy support. The other thing is what we call a win-back, where it's very old systems that are out there. They've got support from somebody else or they're self-supporting. We're going after that business as well. Direct point of sale attach is going very well both for direct and indirect.
Those percentages continue to climb. On the win-back and on the APOS, we're just scratching the surface. We view, and Mark has talked about this, we view that as a very big upside, very big opportunity. I have a team in my group that's part of the renewal sales team that region by region, we have lists of every system that was sold by Sun going back seven years. System by system, customer by customer, we're going back to those customers and trying to get as much of the back support business as we can. We're just really starting to come up that curve. We view it as a big opportunity.
Thanks. I think actually we're out of time for questions. Thank you all very much. Thanks, guys. While we're getting set up on stage, we're going to have Safra come back up for a little bit of Q&A. We'll take it from there. It sounds like we're just about there, and Safra's ready. Why don't you come on up and a little bit more Q&A? Thank you.
Yeah. Hey, guys. I just talked to Larry. I don't know if some of you have flights and things you want to make. Unfortunately, I think he was a little bit too optimistic about his ability to actually get here, of course, due to different sort of family-related things. He's not going to be able to be here. I'm sorry. If you want to make a flight or whatever, I won't be offended at all. You can run for your flight or whatever. I just found out, I would have let you know. You can still make the 4:20 P.M. Otherwise, I'll take questions, and only if you have them. OK, you over here straight in front of me. I just can't see that far.
Hi. Yes. This is Phil Winslow, Credit Suisse. We talked a lot about the changes that you all made with Sun, fixing the product portfolio, the supply chain, and so forth. You all have mentioned the new T-series chip that's coming out. How much do you think that is going to sort of affect the forward revenue, given the fact that it's been so long since we'd had a rev of the T or the M chips? Is there some pent-up demand there?
In fact, I wish it had been so long. This is the first really good, fast chip we've had in a long time, actually. The T3 is not very old, but just not fast enough. We think it's going to be significant, actually. Early orders are looking very, very good, very fast. These systems are not switched over so quickly, and they all need to be qualified. The new chip and the new equipment need to be qualified for the applications, etc . I think it's actually very good, very important. It's an enormous improvement. I'm very, very proud of the team. As you know, when we took over, they were still putting an enormous amount of effort in the Rock microprocessor, which they jettisoned. Just the focus on this system, this thing taped out much, much faster than I think they'd even expected.
The preliminary look at the next chip looks spectacular. We think it's going to be, I think it's going to settle down the base very, very well. I think customers are very, very excited because they don't have to make any changes to their applications. I think that that will actually show up in the numbers definitely in Q4, and will have an impact on the numbers in Q3 and a little bit in Q2. Yeah. Yeah, you here, number two. I just can't see well.
Hi, Safra. Matt Logan with Harris Associates. You guys do a great job running your business, both financially and from a product standpoint. You guys also do a great job at telling us how well you do running the business and from a product standpoint. It's not always easy to figure out what you guys are concerned about, what keeps you up at night, and what we should be thinking about as well. Is it competitors? Is it some disruptive technology that's five years away? Is it cultural risk? Is it execution? I'm just interested in hearing what really worries you longer term.
Luckily, we don't need much sleep. We worry about everything. I mean, I got to worry about everything, Larry, Mark, the same. You don't have to. I'm going to tell you, today is the absolute craziest day because we have our global managers meeting over there, which is like a giant pep rally. Then I come here. I've got to be stayed. It's like going between a strip club and church. I got to just take a minute to get my bearings here. I can't even remember what I told you, what I didn't. I can tell them stuff. I can't tell you. Sorry. OK, back to your question. We worry about absolutely everything. We are enormous, and yet we often change very, very quickly. It is absolutely true. We are in a lot of businesses, but they're all basically completely interrelated.
Luckily, I don't have to worry about video cameras and all sorts of other toys and things that our competitors and others, printers and all sorts of crazy stuff that doesn't matter to me. It's all sort of the same thing. We worry about everything. We are in nonstop. We're like that little Peanuts character that always has a cloud over their head. No matter how perfect we are, we don't start the day looking in the mirror and say, wow, another gorgeous day. We go, oh, what's wrong now? Supply chain, you improve 600 basis points in operating margins in hardware, and you say, what about 1,000? We just don't let up. It is always, what are we doing wrong? How do we get better? Every one of our competitors we take seriously. That's incredibly obvious.
In the database business, we've been so far ahead for so long, it's as if there's no one in the race. Yet, there actually are. We keep coming out with new versions and new products and new things when our competitors, the big ones, Microsoft and IBM, are not doing anything as far as I can tell. Even in our absolute top, most important product, new versions constantly. Thomas just showed you what we're doing with Hadoop and all of this. I know a lot of you are absolutely petrified of big data. We are big data. This can't possibly be any better for us. The old days when only people created transactions, and those are what we kept track of. There's only so many people. They only create so many transactions. Whatever. Now machines create data. It's fantastic.
They create so much data, it's almost impossible to figure out what's important. Not only that, we can help you both figure out what's important. If you want to action on it, to do something with it, you come to us too. This just helps us. All of these trends play very, very well for us. Do we worry about execution? Absolutely. I have more than 105, almost 110,000 employees around the world. I am always at risk of my weakest link. I worry about every single one of our customers. If something's not configured right and they've got an issue, we are absolutely insane over it. OK? Everyone is notified. We go crazy. What do we worry about? Things like that, everything, absolutely everything. Yes?
Hi, Safra. Jason Maynard.
Jason, these lights are totally blinding. I can't even recognize Jason with that hairdo. Sorry.
Thanks. I have a question for you. A lot of your competitors, now that you're in the systems business, integrated systems, talk about attach rate of services. They have big, big offerings. Can you maybe opine on how you think this market plays out over the next two, three years and the multiplier effect of services being sort of the gravy for these guys? What does that mean for customers as you play your strategy forward?
Services really have two sides to it. Obviously, there's the financial side for us because that's really a returning annuity. The true value and what you're going to see more and more focus on, and a number of the guys covered it in their presentations, is that we really want to automate the labor and really automate the service out of service. What happens is we want these systems to be, I think you saw it and some of you saw it over in OpenWorld, self-fixing, self-patching, self-correcting, self-configuring, really almost self-deploying. That is what we will be focusing on for us. That is sort of what we break our heads on in development meetings that Larry holds nearly every day of the week. That's what we're thinking about, those kind of things.
Obviously, from a financial point of view, that totally raises the relationship between us and the customer because the system itself is so central. They can spend less managing it, and yet, we can earn more. That combo is just intuitively perfect. It builds the relationship between us and our customers because their experience is so much better. The better their experience is, the better and more likely they're going to buy more. That's the business we're in. Yes? Go ahead.
Hi. Thanks, Safra. Heather Bellini.
Hey, Heather.
Hi. I was wondering, this was a question I had for Larry. What have you guys learned from Amazon's web services offerings over the last few years? How will you differentiate Oracle Public Cloud? More importantly, given they've been the low-cost provider in the market and looking at their margins, how do we think about your view of the margin potential of that business as you start to achieve scale?
Remember, what theirs is, some things are similar, public cloud. Some things are different. It's our stuff. It's our intellectual property. What they put up is they'll put up some open source, or you can bring your own licenses to it. That's lovely. That's fantastic. With us, you can have everything. You can have the database there perfectly tuned, Java perfectly tuned, a security service. This is a next level. They are simply, really, they give you access to computers plugged in somewhere. It's very nice. We love the way you can self-service, self-provision, all of those things. It's all fantastic. They benefit to some extent from the economies of scale. Unfortunately for them, they do not make hardware. They do not really make the software that people run. We make all of that. We love them as a partner. We love them as a customer.
They run the e-business suite also. They just closed their financials. We keep a very, very close relationship with them. However, our competitive offer is different. The financial metrics of financial capabilities for us are very different because we are Exadata. We can run this on Exalogic. We can do all those things to give us a massive, massive advantage. They can buy those. They do. As a general matter, that's not how they run their cloud. We have all the intellectual property that the customers want to run. If you want to get everything all in one shop, all optimized, all provisioned, et c, you can't do that on Amazon. It's for different purposes. The market is enormous. We don't have to take from each other. As Larry said, we're perfectly happy. You've got your own licenses. You run them over there. You want to move them to us.
You want to move them to yourself on-premise. All works for us. We make out very, very well regardless of where you go. Yeah.
Ross MacMillan from Jefferies. I had a question on Fusion Applications and your public cloud strategy because right now, you're the only company out there that has an enterprise-class set of applications that can be run on public cloud. I guess the question is really, what's the appetite for that? What are you seeing from customers in terms of large customers saying, sure, we're willing to put our general ledger out there or other elements of their core applications? Thanks.
It's kind of, you know, one of those answers that goes, it depends. It depends who you talk to. Some customers are willing to consider it, and they're willing to consider different parts of the cloud. We have hundreds, a couple of hundred early adopters in the early adopters program, and they're good with that. That's great. Some of them are going to bring, they're all on our cloud right now, and some of them will be going back onto their premise when they're live and completely ready. Some of them are staying with us. The others are staying in the cloud with us. A, it depends on the application. B, it depends on the company.
I would have said, frankly, that it depends on the industry, you would think, but yet, really, more than that, it really depends on the culture of the company and their willingness, their different level of, I don't know, it's sort of, it's almost an appetite to do these things. We see it, you know, it's very dependent on the companies. We are encouraged by the interest. A lot of folks are just going to want to upgrade and stay on-premise, and that's OK with us too. We feel like it's important for us to give them the choice. We're not doctrinaire about it at all. They want to be in-house. Even our CRM customers who have had cloud, oftentimes, this is in the cloud. Another part's on-premise. We want to be able to give those customers the choice, and that's working out very well.
Where do you want me to go? Ken's bossing me around. You, over here, I can't see you.
Hi. Walter Pritchard with Citigroup. I'm just wondering, when we were here last year, I think there were some questions around Sun and the $1.5 billion target. You clearly leaped right over that. As we look forward, specifically your measure, or from our perception, you leaped over it. I'm sure it was hard work. As you stand here today, are there specific sort of projects and cost reductions and so forth related to that transaction? Are we basically behind those and now it's more of an overall Oracle leverage and so forth set of things that you're looking at?
You know, it's not as much a cost reduction. It's just a scale in that our big thing is, as you see, we're not a big layoffs or anything like that. You see, our headcount keeps going up. What it is, though, is about making very, very clear decisions, avoiding duplication of effort, and then growing without layering on more cost and getting real leverage. I know I probably used that word like 100 times. I'm sorry. We'll make it a drinking game next time. It really is that. It is economies of scale. I mean, listen, we've been selling about $1 billion of hardware a quarter. I'm supporting a whole hardware sales force that could absolutely sell more than that. We put in development, we've just built a brand new microprocessor, and we've got another one a little ways out. I got to now start selling those.
The development is done. Now it's the time to actually reap the rewards. Will we continue to optimize our operations? Absolutely. We're not at all perfect there. Sun Microsystems gave us a lot of low-hanging fruit. We still find more. There are a lot of things we still have to improve. As we continue on, we've owned them for what, a year and a half? There's still more to do. There's a lot of leverage left. OK. Yes.
Hey, Safra. Brent Thill with UBS. Just on the storage side, there's a lot of friendly vendors that you've gone to market historically. I think you had a few present at the keynote. When you think about how you're balancing your relationship, but also interest in this business, when you're the newbie in storage coming up the stack, can you just talk through how you're balancing that out and how you think about that evolving over time?
First of all, as a general rule, this whole industry is about, I guess, what they call coopetition, competition and cooperation. Pretty much every one of our competitors comes to Oracle OpenWorld. They're all here. If they didn't come, we'd only have each other. It's great having them here. We put them on a nice big stage. EMC, I actually introduced Joe Tucci. He's a good friend of mine. I do an event with NetApp. They are the big guys in storage. We're the little guy. Little by little, we've got things that we think are very, very competitive, whether it's our ability to store a lot more data because of compression or because of speed or whatever, performance. We do have some advantage in that we know the software better. We continue to work with those guys. Obviously, we have sort of a Trojan horse storage, which is Exadata.
The Exadata storage servers, expansion racks, which we don't even show you here, is another, which is the kind of thing you add on to your Exadata. Spark Supercluster also comes with storage. We actually think your storage should come, it's like your fries should come with your burger and a Coke. It's called a happy meal. That's the business we're in often. We're also in the storage business. We are the new guy. I think we compete with pretty much everyone that goes on that stage except us. We expect that to continue. We do think, as I said, we've got some advantages. Some of them are compelling enough to convince our customers to change the way they've done things. Little by little, that's happening. OK. Yeah.
Hey, Safra. Two questions. First, what is the Exadata revenue run rate? Because one of the things that you and Mark were talking about was that we're focusing on the wrong items. In order to focus on the right items, we want to have a basis point, starting point. Second question, with respect to you guys.
I can't wait. If you ask me how old I am, we're finished. OK? Seriously.
With respect to the fact you guys are, you're an executive and a global technology leader, I really would like to get your perspective on the macro environment. What worries you? What doesn't worry you? You have to worry about planning for the future years and allocating all that spend. Thank you.
Sure. OK. Let me answer your second question because we do not actually disclose the first one. It is a very good effort following the rule, don't ask, don't get. Let's talk about the macro environment. As many of you have heard me say like a trillion times on the earnings call, I am not actually an economist. The Wharton School did give me a Bachelor in Science and Economics, but I am not an economist. I have absolutely no idea what the heck is going on over there in the economy. I have no idea if I should be worried about the retirement age in Greece or not. Clearly, at Oracle, we do not take retirement age very seriously. As some of you know, the reality is I think this is sort of as it is. This economy is as it is.
I think for us, we are making the very directed decision to expand because we believe that our products do very well in both great markets and in a crisis situation because our products are extremely compelling from a return on investment point of view. One Exadata, you have taken out maybe a dozen other servers. It ends up being much more performant, much cheaper, much less power, much less labor. We think we have a lot to offer there. Now look, we have all been through a global economic meltdown. I think we did OK under the circumstances. As far as companies, our earnings per share continue to grow. Some people care about earnings. We do at Oracle. As far as could things turn really, really badly? I do not know. If they do, we will make some adjustments, no doubt.
Remember, a lot of us have gone through the internet bubble implosion. We came out of it stronger than ever because we were able to wipe out and acquire some of our competitors during that period. We came out of it with a 15% higher headcount in R&D than when we went in. I will tell you, if the stock market and all this, we got real problems, we are going shopping because if stuff gets cheap, it becomes very compelling for us. We are here to stay. We are not going anywhere. In rough times or whatever, other companies are not going to be nearly as solid as we are. As a result, we are going to get to pick up market share and opportunity. We got a lot of margin points and a lot of money to hold us. Go ahead.
Joel Fishbein from Lazard. Just to follow up on an area that you guys haven't really talked about very much, but it's been in the press in terms of Java, monetizing Java, one of the most pervasively deployed pieces of software out there. Obviously, you've had this success against SAP. Now the court with Google loves to hear what your strategy is in terms of monetizing that from here.
Sure. Obviously, the first way we monetize Java is with Fusion Middleware and Fusion Apps and all that. Java is the absolute central platform of modern computing. There's no question of that. First of all, that's the way we've always monetized Java. Now we have it within Oracle. As you know, we've moved to a new version of Java through the Java community process. It had been stalled for many, many years under Sun Microsystems. We were able, with our good friends IBM and SAP, to move it forward. Every one of our competitors voted for the new spec, actually, except for those Googlers. Now Google, I'm not going to comment too much on litigation. We believe that Android is actually infringing on Java. We are handling that with them and with the courts. As a general matter, Java continues to be strong.
It is very tied to making sure that it is a singular Java shared and monitored by the community process. Companies do continue to license Java from us, as do, of course, all of our competitors. That remains a strong business. Adam.
In the first quarter, which isn't usually a very strong big deal quarter for you all, you had a 40% plus increase in over $3 million deals. An enormous acceleration in very big deals. Just watching the slides today and thinking about what you're doing on the sales side and bringing together all of the pieces of the stack, do you think we've seen a permanent fundamental shift in becoming more strategic, doing larger deals, gaining share of wallet that will suggest the first quarter is the first of many where we see that kind of fundamental transformation?
I mean, I think it's that annoying target metaphor I always use. I mean, people walk in the store for one thing before they know it, their cart is full. They are making a real transformation. There are a lot more of our customers who are, in fact, kind of going for a lot, going for a number of pieces because they understand how these pieces work together. In this past quarter, though, there wasn't any specific. I always hear about that giant deal. It's like the Loch Ness monster. I've never seen it. It is absolutely true that once again, deals are, I used to say this a long time ago, chubbier. That is because they're in there for one thing, they get a little more.
Also, we're spending a lot more time with our customers, sort of showing them really all the strategic benefits of going with us and leaving a lot more of the responsibility with us. Since we are based on open standards, as a result, customers still have choice. They still have an enormous amount of leverage over us. They're signaling that I have to get off stage. Is that right? OK. Can I say thank you? OK. Then you can wrap up. OK. Thank you very, very much. You've made this absolutely delightful. No, really. I'm serious. No, really. It is really nice to see you all. I'm seeing a lot of very familiar faces. I'm very grateful that you're here. I am very sorry about this thing with Larry this time. It just couldn't be helped. I'm very sorry, and he sends his apologies. Anyway, thank you very, very much.
Thank you for your time. Thank you for your support. Thank you for caring about us. Here's Ken. Thank you.