Soltis Exiting IBM, But He's Not Leaving the '400
Published: December 8, 2008
by Alex Woodie
This New Year's Eve, three weeks from now, AS/400 fans around the world will raise a glass and say a toast to Frank Soltis, who is set to retire from IBM after an incredible career spanning five decades. And while there will be some mixing of tears and champagne--especially upon the frozen tundra of the Rochester campus--you don't have to say your goodbyes just yet. Soltis aims to remain active in the System i community, as he explains to The Four Hundred in this exclusive one-on-one interview.
Alex Woodie: So you're leaving at the end of the month. That would be 40 years with IBM?
Dr. FS: Interestingly enough, it's 40 years with System i. Officially, I have 45 years with IBM. I got started with a summer job many, many years ago. During the first five years, I kept taking an educational leave of absence, going to school. But one of the anniversaries I celebrate is November of 1968--this is kind of scary--when my wife and I moved to Rochester permanently. My first assignment was basically on what we call today System i.
AW: What did you work on from 1968 to 1978?
Dr. FS: The System/38. When I walked in the door back in November of '68, my very first assignment was in a group that at that time we called Advanced Development, and my very first assignment was to design the replacement for the System/3. The System/3 hadn't even been announced at that point. It was going to be announced about six months later, in 1969, but the mindset in IBM--in fact the whole industry in those days--is you had to come along with a new product, a whole new system, about every five years. Management here in Rochester was looking at the 1974 to 1975 timeframe to announce the replacement for the System/3, and so it was my role at that time to come up with a design for that particular system.
It turned out that we actually had it delayed, because in the early 1970s we had a recession, a rather severe recession--probably nothing compared to what we're having today--but it was rather severe at the time. As a result, IBM froze hiring, so we couldn't hire any new people to work on this replacement for the System/3, which of course eventually became the System/38. So it got delayed a couple of years before we could start staffing up and actually implement this. That has been the unusual part, in that I have worked on one product, if you will, for my entire career, all 40 years.
AW: Like you said, IBM used to launch an entirely new system every five years or so. Do you think the era of operating systems is passed? Are operating systems obsolete? Are we just stuck with the ones we have?
Dr. FS: No. I don't happen to believe that at all. The concept in those days was really coming out of the 1950s. And it wasn't just an IBM belief, it was really across the industry. If you think about what's going on at that time, it was hardware. Computer companies built hardware. It really wasn't the software product. Granted it certainly was with the System/360 and even some of the things we did with the System/3 products, which of course eventually evolved. . . . The five years was just the wrong number from that standpoint. From a development perspective, there was a strong belief in management that you had better be ready to announce something new in five years. And that's why I ended up with that particular assignment. Obviously what became more important during the 1970s and beyond is really the software.
I happen to be a firm believer that we'll see more operating systems. I don't know how much time we can take to explain that. But I clearly see that there are additional operating systems coming, because, in fact, the hardware in a very real sense is forcing some new operating system environments. Now one of the things we may end up doing is enhancing existing operating systems. But I wouldn't be totally surprised to see some new operating systems--open source operating systems--coming along in the next few years.
AW: So it's more of an evolution, but you maintain that core framework that's here?
Dr. FS: Its power is really on building business systems, and with a business system, you have to protect that application investment. And that's really what the customer is investing in--it's their application. That has been our philosophy here in Rochester all the way back to the System/3, that you had to continue to support that application, and not cause a major disruption in the business.
AW: It could be said that you guys did that better than anybody.
Dr. FS: I think so. But then I'm a little close to it.
AW: So what are you going to do come January 1?
Dr. FS: I'm not necessarily leaving the i world. First, some quick background on the unification of i and p. We actually started that over a year ago, when we formed Power Systems. I had been one of the drivers of that. I had been very much a proponent of merging our systems onto common hardware, and of course sharing common software. So it was back in January when we did a second reorganization in IBM. At that point, the jobs that were specific to i or p suddenly became cross-platform jobs. I said, 'You know, I've been on this one product for 40 years--I really hate to break that record.' And so I was planning to retire back in January, but my goal of course was to stay around with System i folks, and as it turned out, to make a long story short, I was convinced to stay around and help with the transition period.
That's really what I've done this year. I've been traveling, meeting with customer groups, business partners, and so forth, around the world, to talk to them about this unification of i and p, and why it's a good thing, and I firmly believe it is. And so that period is over, so the idea now is to retire at the end of the year. But I'm already talking with various user groups, business partners, and so forth, about continuing to support i and i customers around the world. I'm not sure how long it's going to go on, but that's really my goal.
AW: So we'll see you next spring at COMMON in Reno?
Dr. FS: Yes, you should.
AW: Will you continue to teach at the University of Minnesota?
Dr. FS: Yes. I'm looking forward to being able to do a little more than that. I've been doing an awful lot of traveling. This year was very hectic travel. So I had great difficulty teaching a regular class. I've been teaching seminars . . . . But a weekly class, I had great difficulty doing that. I'm hoping to get back to that next year and in the future.
AW: Who's going to take your place at IBM as the Grandfather of the AS/400?
Dr. FS: [Laughing.] Like I said …we have nobody who's focused exclusively on i or exclusively on p. We just don't have that role anymore. That job went away essentially. That's the reason I decided to retire.
AW: What do you guys call the platform up in Rochester these days? Is it Power Systems, System, i5 . . . ?
Dr. FS: I'm going to laugh a little bit because if you were to walk through the halls in Rochester, you probably hear them talking about the AS/400. Just like you do in the whole community. But what we're talking about is the Power Systems is really the hardware. And what we've done is separate the operating environments, so IBM i is the operating environment, which most people think of as an operating system, but in fact it is really a lot more than just an operating system. But IBM i is the term we use for it. And of course running on Power Systems also are Linux and AIX.
It gets a little bit confusing in the sense that, years ago when I could walk into a customer shop, and there would be a System/36 customer, and they would have only System/36, they'd have some terminals around that were maybe running Windows. But they were a System/36 customer, and you could talk about as being 36. Today, that clearly is not true. I've spent quite a bit of time the last couple of years with some of our largest customers around the world, and you walk in, and they're using i, they're using AIX, they're using Linux, they're using mainframe, they're using everything. And so when these systems are capable of running these multiple environments, it's difficult to describe them as a System i or a System p, and that's the reason we just generally talk about them as Power Systems.
AW: How do you see IBM i and AIX co-existing? Do you think they're complementary to each other with the applications they bring and managing the systems?
Dr. FS: Absolutely. One of the things I used to joke here over the last couple of years, I thought we should re-announce some of the things we had done in the past.
Back in 1998, we announced an application environment called PASE. Remember PASE? PASE is AIX. It's the runtime environment for AIX. Internally we have a bit of a joke about it. I was involved in that. The name that was originally given to it was Private Address Space Entombment. And the reason we called it private address is because in the Unix world, the address spaces are separate for the process, they're private to the process. And in the AS/400-i world, everything is in single level store, it's all shared. So what we had to do is we had to take within the single level store, and carve up a number of these private address spaces, and therefore we could implement a pure Unix. When we came to announce it, they said the world won't understand Private Address Space Environment, so we'll change it, and I think they changed it to something like Portable Application Solution Environment, which I have no idea what that means. They kept the acronym PASE, which is not a real word anyway.
But if you look at what it is, it's AIX. And so over the years, we developed code, some middleware product, that runs in AIX, we can run it in the so-called PASE environment, and it's all part of the i operating system, it's fundamental to it. So the two of them have been coming closer and closer and closer together for more than 10 years, so it just absolutely made sense to have one system.
AW: But from a cultural standpoint, do you see any tension between the p system guys and the i side?
Dr. FS: Culturally, there's a big difference, yes I agree with that. When you come to the briefing center here in Rochester, the discussion is mostly about business. It may be about specific applications, you get into discussions about backup and recovery, all the things you do with a business.
Have you ever been to Austin's briefing center? If you ever have a chance to go there, they have an excellent briefing center. But the focus will be on the hardware. In other words, you'll hear a great deal about feeds and speeds and size of caches and memory bus interfaces, etc, etc. which you would just never hear about in Rochester. It really plays to the different audiences.
When I went to see a Unix customer--a week and a half ago I was up in Canada--there was an AIX gathering, and I spoke to an AIX group there, and my focus then certainly shifted over to be much more on technology, because that's what they wanted to know about. The customer groups are different. But from the standpoint of the system itself, and it's capability to do either or, and in fact the two in combination, that has never really been an issue. But they are culturally different groups.
AW: How well is the Power System i platform positioned today to compete?
Dr. FS: As I sit back and I look at IBM, I see we have two major operating systems--I like to call them operating system environments, because it's really bigger than just the operating system--but we have two that are geared toward business computing. One is i, and the other is mainframe. And if, for example, I'm in finance, I'm running a big worldwide bank or some such thing, I'm going to be picking one of those two platforms. If you go out and you look at those two industries, that's what you find. That's not to say you can't run a business with Windows or I can't run a business with Unix, but there's a different approach to it. It's much more of a do-it-yourself approach, as opposed to the integrated, complete solution. So, to me, it's positioned very well from a business standpoint.
AW: But in terms of the platform as a whole. There's a lot of people worried about declining revenues. IBM is not breaking out the i portion of Power Systems revenue, so all anybody hears is the 82 percent revenue decline with the legacy System i boxes.
Dr. FS: We've been pretty upset with that. I don't like the way it's reported. But … this is not a topic that I know a whole lot about. From what I understand, the reporting is really the combined systems, the Power6. Anything that we sell--i, p, whatever--with Power 6, this year has been reported under System p. Somebody made that decision, I don't know who. I don't think it was a good decision. So when you see these declines, it's really a previous generation, which for the most part we're not selling any more.
AW: If you were to create an entirely new server platform today, what would it look like?
Dr. FS: You sound like my kids. My sons keep asking me when I'm going to work on a second product at IBM. I like to look at design. When I'm teaching at the university--I teach design classes, processor design classes, and so forth--what I like to do with my students is, first of all, lay out my objectives. What am I trying to accomplish? What kind of environment are things going into? What sort of applications are going to be running? We typically look at the design, and then at the end, do an analysis, how well did the designers of a given platform do. And so if I were looking at this . . . [from the standpoint] of a business computer, the general objectives that we have, and which are pretty much shared with mainframe and System i, is I think the right choice.
And I think the fact that something has been able to last 30 years, is because we held true to those objectives. We don't force changes. We develop something that is today what we would call a virtual machine, but back in the early days, we just said it was a Technology Independent Machine Interface. The idea that hardware will evolve very quickly, and we're going to make changes to the hardware, but those changes do not come back and impact applications or customer software. Realistically, if you look around the world, there are very few products that have ever been able to do that, and there are very few products that have had the life of these products, IBM mainframe or System i. So from the standpoint of design, I think the design parameters are extremely important, and they are as important today as they were 40 years ago.
AW: So it would look a lot like the System i today?
Dr. FS: It would have a lot of the same characteristics as the System i does today. From a user perspective, from a customer standpoint, it would have a lot of those same characteristics. Details would change of course--you never do everything the same way.
AW: Looking back over the last 20 or 30 years, is there anything you would have done differently?
Dr. FS: Every design is a compromise. I don't care what the design is, or what you're designing, there's a compromise in there. And you know based on some things that you know later on, you probably would have made different decisions. But I think the overall concept would still be the same.
AW: Can you give me an example of something you may have done a little differently?
Dr. FS: Several things. One, I got involved with the design of Power back in the early 1990s, I led the team at IBM that designed the architecture for what is today the Power architecture. I don't do chip design, but I've been involved with architecture all along. At that time, we had the view that Power could be the only hardware platform that IBM would have. My goal at the time, and in fact the goal of a number of people at IBM at the time, was to come up with one platform that could be used across all of our systems. We haven't quite achieved that.
You may recall back in the mid-1990s, we had a PC operating system, OS/2, which is still used heavily in the finance industry, banks and insurance companies. But we obviously lost the battle to Windows. But back in those days the idea was OS/2 was also going to be a very strong platform for us, running on Power. Not that I was necessarily directly involved in OS/2, or trying to market OS/2, but ideally if you could go back and re-do that, I know IBM would do it differently, we would have really promoted OS/2 a whole lot more. That's one of those disappointments from that standpoint.
I was disappointed a couple of years ago when the mainframe decided to go their own way with their own hardware, as opposed to using Power technology. We, of course, are using it in systems, the Power Systems, but also in storage and supercomputers. We're selling it outside in game consoles and that sort of thing. But the vision that many of us had back in the early 1990s was this would be the one platform. It'd be used across IBM, and by the way, it would be used outside of IBM extensively. Some of those things didn't happen. I'm not sure if going back in time if there were some of us on the technical side that maybe could have influenced that. I really don't know.
AW: Do you think there was more opportunity to get more operating systems on Power, maybe Windows?
Dr. FS: Yes. But again, my focus back in those days was upon AS/400. So it wasn't my decision to say which ones we go after. In hindsight, if we could do that over, and I suspect there are others in IBM who would agree with me, we would more aggressively go out after more operating systems.
AW: Do you think there's too much hype in the industry these days? Have we gotten too far away from business requirements?
Dr. FS: Of course. Isn't that true with just about everything? You've got to have something that's really sexy to get people excited about it. Yes, there's a lot of hype out there. But when it finally comes down to it, if you really sit down with almost any business, the final decision really is a business decision.
AW: How has IBM changed in the 45 years you've been there?
Dr. FS: One of the things I absolutely love about IBM is we change, and we change rapidly.
When I was more involved in the day-to-day laboratory operations, we would have new hires coming in, and I would always set up a session with our new hires within six months of their being hired. Within six months, there would be a lot of changes. Somebody hires into IBM, and it's a certain way, and all of a sudden everything around them seems to be changing. So after about six months, it's fun to sit down with some of the new hires and say, 'Well what do you think?'
Change can be disruptive to a lot of people. A lot of people say, 'Oh my gosh, it's changing, what does that mean?' and to me that has always been the positive aspect of IBM, is that we will change. We're in a very dynamic industry. Things change all around us. Technology changes daily. And you have to be able to move with it. And so to me that has always been the positive part of working for IBM. But it takes some getting used to. And like I said, I would meet with some of the new hires, and would tell them 'This is the way we operate. This is good news. When technology changes, we will change to adopt that technology.'
AW: How has Rochester changed?
Dr. FS: Rochester probably has not changed as much. You've got remember, we're all Midwesterners. I grew up in Minnesota. This is hard to believe, I know, but I can't imagine living any place else. Like I said, it's kind of harsh for somebody in Southern California to understand. But if you look around at the people in Rochester, most of us are from this part of the country. For years, we were trying to recruit people from either the East Coast or the West Coast universities, and with not very much success. Think about it--somebody graduates for the University of Southern California, and we say 'We've got this wonderful job, but it's in Rochester, Minnesota.' It just doesn't happen.
If you look at the people around here, they come from mostly Midwestern universities--Minnesota, Wisconsin, Iowa, Illinois, Indiana, the Dakotas, and so forth. These are the places that these people who work here come from. And so people living in Rochester are here because they want to be here, because it's home.
The other thing that happens, you don't step out of IBM Rochester and walk down the street and go to work for somebody else. That's not totally true, we've got other businesses here, and business partners. But it's not quite the same as Austin, Texas, or Palo Alto.
And so as a result we have a very stable workforce. I joke about working for IBM for 45 years, and never really being outside of Rochester or moved outside of Rochester. I always refused whenever something came up to move, but I'm not alone. There's an awful lot of people that start here and retire here. So you have that very stable workforce. And we all have pretty much the same values in our design. And I think that in my mind is the reason that Rochester products have been so successful.
AW: What are you going to miss the most about not working at IBM Rochester?
Dr. FS: Well, I still plan to stop out there. People, obviously. Being around the people. Hopefully, I can still stay around some of the customers, business partners, some of the user groups. That's really what I'm looking forward to is maybe I can spend a little more time with them than I've been able to do in the last few years because of other responsibilities at IBM.
AW: So you will be doing more platform-evangelism type stuff?
Dr. FS: Yeah, I'll be doing some of that. One of the things I'd love to do is get more involved with COMMON and some of the user groups, which really as an IBM employee I wasn't totally able to do. We obviously have IBM liaisons working with the user groups, but that's a very specific responsibility. If I was more independent, I might be able to participate a little more with some of the user group activities, and help with some of that direction.
AW: Can you talk about anything concrete that you have planned?
Dr. FS: Not yet. I'm still in the process of negotiating on some of these things right now, as to what I'm going to be doing. Like I said, I've been approached by a couple of different user groups already as far as getting involved with them and their conferences. I've spoken to a couple of business partners that have conferences. I'm in the process right now of going through all the paperwork. I didn't realize how difficult it is to retire in this world, getting all the forms signed and everything. There's more paperwork involved than it would seem necessary.
But it's been a great ride. It's something that I never anticipated. I never figured this would be the kind of job that I have. I just kind of fell into it. But it's been fabulous.
Post this story to del.icio.us
Post this story to Digg
Post this story to Slashdot