Remarks of Frank Soltis, Chief iSeries Scientist

iSeries Town Hall Meeting, Fall COMMON, September 18, 2005

The reason we had that long introduction is as I'm getting older it takes for me to get up here. So good to be back in Orlando. I do have some mixed emotions though. One of the things that we've started with the family--as my wife calls it a tradition--is whenever we come to Orlando we bring one of our grandchildren along. And sure enough on this trip we have one of our grandchildren along. And so earlier today, I said to my wife, "Why don't you come along, listen." The answer was as expected. She said no. So they're off at Disneyworld enjoying themselves.

Something happened to me, several weeks ago at Mall of America. I was there with my brother-in-law, we were wandering around the Mall of America, doing what most husbands do on Saturday afternoons at the Mall of America--waiting for our wives, right? And as we were walking along, I saw another couple there, and I recognized the woman in the couple as somebody who had worked on an iSeries because I recognized her from some of the user groups in the Twin Cities area. So we stopped and we chatted for a bit.

During the conversation, my brother in law asked her, "What do you do, what's your job?" And she said, "Oh, I'm an RPG programmer." And he said, "Wow, that's pretty exciting. You must really love your job." And she said, "Well, yes, yes I do. I really love my job." I was taken a little bit. I like RPG programmers, but exciting? Whatever. So we chatted for a couple more minutes and walked on.

So I said to my brother-in-law, who's not in computers, doesn't work for IBM or anything, "Why did you react that it was a real exciting job being an RPG programmer?" And he looked at me kind of funny and he said, "In many circles, RPG programmers are viewed as geek gods." Geek gods?" I don't mean to insult anybody in this room, but most of us don't think of RPG programmers as geek gods, right? So I looked at my brother-in-law and I said, "Where did you hear that?" And he said, "Well it's on Slashdot all the time." Now my brother-in-law is a big Slashdot fan. Most of don't get involved in that. And I said, "Slashdot says RPG programmers are geek gods?" And he said, "Yes." I said, "My gosh. I didn't realize anybody at Slashdot even knew what Report Program Generator was." At which point my brother-in-law looked at me kind of funny and said, "What's Report Program Generator?" I said, "Well it's RPG, Report Program Generator!" And he said, "No it's not. Roll Playing Games is RPG." "Roll Playing Games?," I asked. "Yeah," he said, "everybody knows that!"

Well maybe not everybody.

So when I got home that night I went to my ultimate source of information--Google--and put in RPG. And guess what? Roll Playing Games is what RPG stands for. In fact, I had to go through 20 pages before I found something that related to RPG as we know it. There's more RPG sites out there than you can count, and they're all Roll Playing Games, and yes, game programmers are geek gods, they certainly are.

So it got me to thinking a little bit. I thought, how can we turn our RPG programmers into geek gods? Not a bad thing to do, right? And so I came up with a crazy idea, just one of these absolutely nutty ideas. Well, you know, all iSeries are obviously based on the Power technology. And every game console is also based on Power. The next generation from Microsoft is called the Xbox 360. Don't you love that name? I think somebody else has used that name. Yes, the Xbox 360 is coming out later this year in time for Christmas in time for Power and it's essentially the same technology that's used in the iSeries. Nintendo has been Power for a number of years. And next spring sometime Sony is coming out with the next generation of PlayStation that is also Power based. So if you look at it, all the game consoles in the world--and by the way, is that a big business? Yes it's huge--they're looking at this next generation of game consoles and planning to sell somewhere in the vicinity of 200 million of these things. And they're all based on Power. And so I say, logically, we can run a game on an iSeries, right? And why don't we, when we ship an iSeries, ship a couple of games on it?

[Much applause from the COMMON crowd.]

Nobody's going to buy an iSeries because it has the games, but lets have a little fun with it. So obviously this is a pretty wacky idea for IBM, so I needed to bounce this idea off somebody else who was equally wacky. So fortunately I don't have to go very far; Malcolm Haines is usually at hand. And so I bounced this idea off Malcolm, and he says, "That's a great idea, let's do it." And I say, "We've got to find some forum to show an iSeries game. What about COMMON?" And so we started an effort to build the first video game that runs on an iSeries and demonstrate it right here at COMMON. So I talked to Bev and others at COMMON, and asked them, "What do you think about this? Is this a good idea?" And they said, "Great let's do it!" I had to look around for some technology--I didn't want to drag a big iSeries up here on stage. Fortunately, a few months ago I updated my technology, and of course when I update my technology--you know, my technology--as most people recognize Apple--by the way, there's more IBM technology in Apple than there is in some Intel-based thing--and we're not in the PC business anymore, right? IBM got out of the PC business, so we can buy the best, and the best of course is Apple. Over the years, many of you will notice I never run visuals because I run everything on Apple and it was embarrassing.

So I got the latest version here of my iBook. The only models I ever buy start with i. and I already had it, because I needed it to download my iPod, and of course I've got the latest version of iTunes, for the same reason you do. You're all tuned into the new podcastaways.com site? Isn't that great. You're all tuned into podcastaways.com right? And you've downloaded it? All four of you. Well let me put it this way: You're going to hear more about podcastaways.com during this conference. An iSeries person first of all should be hearing podcasts from podcastaways.com. My favorite part of podcastaways is of course "The Podfather." You know it's like one of these ongoing sagas, whether I watch a TV saga. Aren't you all curious to know what's going to happen in Rome? I guess you don't watch that either. Both of us. When I listen to one of these sagas like The Podfather, I usually relate to one of these characters, and the character I relate to on The Podfather is Franco. He's a bit of a hot head, but somehow I just feel some sort of affinity toward this individual. So you're going to hear a lot more about podcastaways and The Podfather and a few other things during this particular conference.

I have the technology, and I'm ready with my Apple computer to show you the first video game that runs on an iSeries. Who would like to first video game that runs on an iSeries?

[Applause.]

Unfortunately, so would I. I got the word a couple of days ago that it wouldn't be ready in time. I was tempted to call Bev and say, "Can we delay COMMON for a couple of weeks?" But I didn't think that was good news, and I hate to say this, I really do. I'm not one to spread rumors, as most of you who know me. But there's a rumor, and I'm sure it's a totally unfounded rumor, that Malcolm sold it to Microsoft. In fact some people were commenting he drove up the other day in a Mercedes. But we're still working on it, and I guess you're just going to have to come back to the next COMMON to see the latest video games.

So I was very disappointed that I don't have a video game to show you. So what's the big deal about video games? You've heard a lot of us within IBM, read a lot of articles and so forth on this whole world of video games and video game consoles and tat sort of things. I get a lot of customers saying, What's the big deal here? You're in the business of developing business computers, and you're talking about video games. Part of the reason we like to talk about video game consoles is that much of the technology that you're seeing or will be seeing over the next few months here tells you where we 're going in the next let's say five years. In fact, a lot of that technology will make its way into commercial systems like the iSeries.

Incidentally, one of the things that IBM loves to do is five-year plans. IBM is very big on five-year plans. I don't know about your company, but we like to do a five-year plan. And not only do a five-year plan, but we tell you about it, which is really really nice. In fact, I've always thought that was really a good thing that IBM does, is we explain to you exactly where we'll be going over the next five years. Unfortunately, when you hear the presentations about where we're going over the next five years, it sometimes is not crystal clear what we're doing. Let me give you an example.

Back in October of 2000, we announced what we were going to be doing for the next five years. That was our five-year plan. October of 2000. Anybody remember that? Don't all raise your hands at once. Well, unfortunately, what happened is that most people thinking back to October 2000 remember only one thing: We changed the name. Remember that? We changed it from AS/400 to iSeries. And we introduced this concept of eServer. And so a lot of people said, "We understand that part of it when IBM announced it, but we didn't really understand the rest of it." Well if you go back, and you take a look at that announcement, we told you exactly what we were going to do with our various servers over a five-year period. In fact, one of the things we said is we're going to begin share technology. We were going to share hardware technology and software technology across each of the products.

What was so terribly radical about that was the fact that we'd never done that before. The IBM Company for years and years never shared anything from one server to another. And many of you had heard me make this comment before, but it really had with Tom Watson Sr. [IBM's founder], who believed that various development laboratories should compete with one another to develop the next generation of product. And, in fact, IBM was really set up on that basis. So when they saw a need for a new product in the marketplace, at least two of the laboratories would be given the option to come up with a proposal, what they proposed for a product. And there would be an internal judging of the products, and one of those would actually go out to the marketplace. And so there was always this competitive spirit within IBM and in fact one of the things that happened was, as you can logically assume, is we would not cooperate very well with one another. We didn't share our technology with our various locations.

In fact, the term "Fortress Rochester" was often used to refer to the Rochester Laboratory and it was a fortress because we shared absolutely nothing with anybody else in IBM. And in fact, I've made the comment before, if you go back ten years, and you look at the four IBM servers that we had, and you tried to identify any common technology through all of them, about the only thing you could point to was the power supply. We did use some common parts in our power supply. That was it. There was no hardware, no software, nothing else, that we shared across all of the technology. Incidentally, one of the sessions that I was doing here just a couple of years ago, in the audience, somebody raised their hand in the back, and said, "Oh, you forgot about something that was shared across all four products." I said, "Really? What was that?" And she said, "Well, they are all painted black?" And I said, "Yeah, but did you realize we used four different black paints?" In fact, in IBM, it may still be the case, I haven't checked it recently, there were 24 shades of black. IBM has identified 24 different shades of black. Think about that.

So in October of 2000, we said, we're going to change that. We're actually going to start sharing some of this stuff. Look at what we've done. Take a look at the hardware technology that we use today. The iSeries and pSeries, and an awful lot of our zSeries, uses the same hardware. We have brought together common hardware across all of these platforms. Now the hardware is just a small piece of it. The very next step is to be able to share some of the software. Now when we talk about software, what kind of software do we do in IBM? We do systems software, operating systems and so forth, as well as middleware, things like databases and so forth. And so we've begun to bring those together. If you think about the new Power Hypervisor.

By the way, isn't' that an ugly name? It's something that sits under the operating system. Internally, most of us refer to it as underware. Somebody within IBM decided that maybe was not a good name to call it, so we call it the Power Hypervisor.

But it's common code that sits today under three operating systems. i5/OS--remember that one? It used to be something else. AIX and Linux. And you'll see more and more of that happening.

What about some of the middleware kind of products? The Virtualization Engine. That's another one of my favorite ugly names. Great products, but I just don't get the names.

If you think about we did with Virtualization Engine, it's not only the low-level of software, but things like middleware. We took pieces from our various servers, and we're making it common across those things.

So in October of 2000, that's what we announced. We're going to use common components, common tools, and, by the way, we're also going to put more automation into it. Only we didn't call it automation--we called it autonomic computing. Another one of those. I don't know who stays up at night and thinks these things up. Whatever. But clearly that's what we've been doing.

From October 2000 till today is basically five years, right? So our five-year plan is up. And so we did a new five-year plan. And, by the way, we need to announce to you a new five-year plan, right? And we did. At the end of July, we announced the next five year plan for IBM systems, including the iSeries. What did you think of it?

[Laughter from the crowd, since IBM didn't actually do this.]

Pretty exciting, huh? We told you where we're doing with the iSeries for the next five years. You don't care about the future? It doesn't matter? Well, guess what, you're going to hear more about that in the next week, too. We call it the IBM Systems Agenda. And I won't give you a whole lot of the details, but basically it says in the future, IBM systems will be integrated, virtualized, and based on open standards. There's a little more to it than that, but fundamentally those are the three things. So over the next five years, you're going to see us shipping systems--anybody notice that they're not called servers any more? Because eServer was the first five-year plan. We don't need servers anymore. We call them systems, and they are integrated, virtualized, and based on open standards.

Can anybody think of another product from the IBM Company that for example is integrated? We're trying to find anything that is integrated, where the hardware and software all works together, it's packaged together. Can you think of another product that maybe fits that? What does the "i" stand for again, I can never remember.

Virtualized is another of my favorite terms. Did you realize that the iSeries was the world's first fully virtualized machine? In fact, it's never been based on hardware, it's always been based on software. But years and years ago, we didn't know what to call it. Some of you may even remember we talked about a high-level machine interface, and remember this ugly thing, Technology Independent Machine Interface? Anybody remember that? Well today that's virtualization, that's exactly what it is. I can adjust this machine. In fact, when you're dealing with an iSeries or an AS/400 or even a System/38 before that, you're dealing with things like logical devices. They're not real physical devices, they're logical. That means they're virtualized. So for the past three decades we've been shipping a virtualized machine. So wait a minute, I have an iSeries that's integrated, I have an iSeries that's virtualized. What have we been doing over the last decades as far as open standards? We fully embrace open standards in the iSeries world and are continuing to push in that particular direction.

So wait a minute, don't I already have a machine that fits that? And the answer is yes. And in fact the message--and like I say, you're going to hear more details about it this week--is that in the future, we're going to essentially be using the iSeries as a model for our future systems, for all of them.

[Applause]

This is great news, right? Now why did the IBM Company say, "Well, gee, these things need to be virtualized, these things need to be integrated, et cetera et cetera?" I mean obviously they didn't look at the success of the iSeries. Well, yeah, to some extent, but it's the way the world's going. And you can first see it if you look at some of the game boxes.

You see, one of the things that's happening is we're able to package quite a few things on the hardware, including some of the things from the software. So today, when we actually even design a processor chip, we don't design it in a vacuum. Years ago, we used to be able to put a bunch of engineers in a room, and say come up with a new processor technology, whatever it is, they'd build a chip, throw it over the wall to the operating system people, and say do something with it. I think that still occurs in places like Intel. But whatever. And then, once they play with it a little with the operating system, they say, we should have some applications on here, and so they give it to all the application designers, and they say, oh my gosh, what did you do with this? We don't do that anymore. We really have to build these things in a system approach, we really have to give people more knowledgeable software, application, operating systems, everything, tied in with the hardware designers, the chip designers, to come up with these things that are fully integrated. And that's exactly what we're starting to see in the game console world.

Let me give you an example. My favorite chip out there is the so-called Cell chip. The Cell chip is the one that is jointly designed by IBM, Sony, and Toshiba. This started back in about 2001, this effort to come up with a new chip that's going to be used in the Sony PlayStation. As I said, next spring, PlayStation 3 will have this new chip in it. The goal of the Cell chip was to create something that was 1,000 times more powerful than what is in Sony's PlayStation 2. Imagine the kind of video experience you can get with something that has 1,000 times the power, and that's clearly what Sony wants to do with this thing. Well, the chip itself is fairly simple to describe. It has a Power processor on it, and it's got eight special purpose processors on there specifically geared toward things like visual kinds of activities, visual processing. So you can mentally think of this as a central processor and eight IOPs. Anybody recognize the model? Where do you suppose all the interfaces and so forth for that structure were designed, as far as the interfaces for those IOPs, they don't call them IOPs, but those special processors? Where do you suppose the technology expertise exists in the IBM Company? Rochester, Minnesota. I often point out to audiences that, if the answer were Poughkeepsie, I wouldn't bring it up. As long as it's Rochester, yeah, OK.

Also in Rochester, we also build the world's biggest supercomputers, the Blue Genes. In fact, the one we're kind of excited to see put together in pieces is the one going to the Lawrence Livermore National Laboratory, this Department of Energy laboratory. And when it's fully installed---it's going in multiple racks at a time. Every time we put a new piece in, it takes all the records as far as the highest performance processor in the world, and when it finally is in there, it will have something like a million processors, and it will represent somewhere from a third and a half of all the supercomputing power in the world in that one machine, which is pretty nice, right?

It has a couple of interesting characteristics. Obviously, it's all based on Power--of course all of these are. But one of the tendencies of this is when we put a million processors together, there's a likelihood of a failure of a processor every roughly four days. Just statistically that's going to happen. Now think about your own business. Suppose a processor fails every four days. Probably not bad if the applications run less than four days. But unfortunately, some of the Blue Gene applications run for more than a year, so we really can't have that. So some of the technology that we build into these things--and part of the reason that we do these things at Rochester and other places at IBM--is to learn how to solve these problems. So we have built in what is essentially self-healing. There are additional processors in there that take over if one of these fails, so it's invisible to the software. So from the standpoint of the application, it appears that it's always running. Now, wouldn't you like to have self-healing processors in your iSeries? Yeah, of course. And so these are the kinds of technologies that we'll bring out in the future.

So we're starting to look at a lot of these strange things, like the Cell chip. Well, most of us feel that this particular Cell chip--by the way, the reason it's called Cell is you can put multiple of these things together to create a complete systems--is this will be the basis of our hardware technology for our products probably starting in the 2008 to 2009 timeframe. And we feel this stuff is really pretty awesome. It's some of the highest performance stuff you're ever going to see out there, but also has all of the reliability that goes along with that. So a lot of the characteristics that we've all come to know and love with an iSeries, as far as high availability and reliability and so forth, we're investing very heavily in those kinds of things. Certainly the game industry in a very real sense is leading the way.

Also some of the visual capabilities. If you think about the user interface that you and I have today, it's typically a Windows screen--well, some of us have a little better one here--but it's typically a Windows screen, that's kind of the user interface into the system for most of us. Well, that's pretty stagnant. It also came out of all the game consoles some years ago, that was really a game interface. And so if you think about the sort of things they're doing with the next generation of gaming, you essentially put yourself into the game. In fact, today there's something like 200 million people who play online games, interactive games, with one another. And so you start thinking about things like collaborative computing, and clearly that's something that's very very important in the business world. Now, it's a little bit different than in the gaming world. In the gaming world, I can actually go into a game, stand on a hillside, and watch some other player kill a dragon. Maybe that wouldn't be so bad in the business world, come to think of it.

Well, a lot of these technologies we're seeing in the gaming world have application in the business world. That's one of the reason that most of us, as we look to the future, and we talk about this concept of being able to integrate these things and to make everything virtualized, because virtualized, I've got to be able to change the characteristics. How about self-morphing hardware? One of the things that would really be cool is, if you've got all this hardware on a chip, and the user loaded a particular application, and then the hardware reconfigures itself for that particular application. None of this is too far out for some of the things we're looking at.

So the IBM Company has invested very heavily obviously in the gaming industry because we really feel that is going to be the future for a lot of these systems. And we announced that, as I said, in July, as our IBM Systems Agenda. I'm just surprised you didn't understand that. Of course we do have a few folks who are known as lawyers who sort of change the words a little bit, so that when we're finished none of us knew what we announced either. You're going to hear more about that this week.

So the outlook is bright. We're very excited as far as the iSeries is concerned. It's not just a business proposition of 10 percent growth, things like that. The iSeries is right in the center today of IBM's future plans. If you take a look at future technology, future systems that you're going to be buying from IBM from all of our different product lines, you're going to start recognize a lot of the capabilities and characteristics of what you known in an AS/400 and you know in the iSeries. And I think that's good news. So thanks very much.