Dr. Frank Reminisces on 35th Anniversary, Looks to Platform’s Future
June 28, 2023 Alex Woodie
The IBM i community came out in force last week for the 35th anniversary of the creation of the AS/400 back in 1988. IBM corralled a handful of IBM i authorities to speak during three webcasts it held for fans around the world. During a call with European community members, Frank Soltis, who developed the architecture for the S/38 and AS/400, kicked things off by declaring there was really just one platform.
“I think one of the distinct distinctions that I have is that I worked for IBM for 45 years before retiring,” Soltis, the retired IBM chief scientist told host Ian Jarman. “During that 45 years, I worked on one and only one platform. So I think that says a great deal about the IBM i and its longevity.”
If the S/38 and the AS/400 were really the same platform, then should we be celebrating a different anniversary, asked Jarman, who joined IBM in 1983, worked on the S/36 and S/38 before becoming the AS/400 product manager, and currently is CTO of IBM Technology Expert Lab (formerly Lab Services) at the Rochester, Minnesota lab.
“Well, the S/38 was somewhat of a niche product,” said Soltis, who retired in 2008 after 40 years of full-time work with the company (he worked summers with Big Blue from 1963 until 1968). “It was so different that a lot of people didn’t know what it was. In fact, IBM didn’t quite know how to sell it because it was not a System/3 follow-on as they sort of expected, which the ‘36 was.”
The launch of the AS/400 on June 21, 1988, marked the merger, of sorts, of the S/36 and the S/38. While the two systems shared similarity in terms of origin (they both came out of the Rochester lab) and names (they’re off by only two digits), they differed remarkably in terms of market success and technological underpinnings, according to Soltis.
“So the successful one from Rochester from the standpoint of number of applications, number of users, number system sold, was the ’36,” Soltis said. “But the ‘36, like many other systems of this time, had some limitations as far as growth. You brought up the DEC VAX situation, of how they had difficulty moving to a new platform, to a new 64-bit platform, and the ‘36 would have had some of those similar problems. However, with the base of the S/38, the ‘36 with its massive number of applications, it’s ease of use, was really key to the development of the AS/400.”
The AS/400 would trounce both the S/36 and the S/38 in terms of market success, of course. Sales of the platform soared out of the gate, and at its peak within a few years, the AS/400 was being run by 110,000 organizations around the world, pulling in billions in revenue for IBM.
The size of the installed base has been pared down over time, but as last week’s festivities indicate, the platform is still heavily relied upon by business, and much-loved by the people who work on it. The remarkable run really shows no signs of letting up, in what has to be considered one of the greatest success stories in business computing history.
It’s almost amazing to think that this followed some script, but according to Soltis, is basically has. That’s because one of the endearing characteristics of the AS/400 architecture that Soltis spearheaded at the Rochester lab is resilience to change.
When the AS/400 launched in 1988, it utilized a 48-bit addressable memory space. Seven years later, the platform would shift from a CISC to a RISC architecture, introducing 64-bit computing to its customers. The capability to rip out the underlying hardware without impacting the software sitting on top of it owes to a nifty bit of tech that IBM called the Technology Independent Machine Interface, or TIMI.
“From the very beginning, the goal was to make the system last for a very, very long time,” Soltis said. “It was pointed out to me that, during its lifetime, the IBM i has now has had something like 25 different processor technologies implemented, 25 different implementations – something that’s almost impossible with other platforms.
Today, the IBM i is firmly established on the IBM Power processor. It would be hard to imagine an IBM i server today that didn’t run on a Power-based Power server. But thanks to the tremendous technological flexibility that Soltis and his colleagues built into the architecture, a Power future wasn’t necessarily foretold from the beginning.
“I also believe that the key development of the IBM I was the merger with Power processor,” Soltis said. “IBM Rochester engineers wanted to build their own processor, as you could expect. And I’m sure it would have been a wonderful processor, but it would have been a different one. And the corporation, headed by the president of IBM – we had a president in those days – he came to Rochester and said ‘I want you to use the Power processor.’
“And so I put together a small group of people and we outlined all of the things that we needed in the Power processor,” Soltis continued. “You don’t use floating point processing in business applications, right. So we needed that. We needed single-level storage and so forth, and we worked with research, who owned Power architecture as well as IBM in Austin to have those incorporated and over the years those were added to the processors. . . We’re running on Power and that will guarantee a long life for this platform.”
Of course, just because IBM i runs comfortably atop Power today – and likely for many years to come – it doesn’t mean that this will always be the case. Thanks to TIMI, the platform is adaptable to changing hardware. And Soltis sees lots of change on the horizon when it comes to hardware.
“If I look at some of the technologies that are on the horizon, obviously the interest in AI, of some of the new technologies to possibly replace semiconductor technology in the future,” he said. “IBM has announced that they’re working very heavily on things like quantum computing, which over the next several years we expect to maybe take the place of some of the traditional semiconductor systems that we have today.
“What we have been able to demonstrate with the IBM i and its predecessor is that we can easily incorporate those new technologies,” he continued. “You can’t say that about an awful lot of systems in this world, but the ability to be able to incorporate those technologies means that we can move with the requirements of our customers, and also with the requirements of the new technologies of the future.”