Paragon Software International, Inc.

Big Blue's tortoise pips hares at the finish
Tuesday 11 July 2000

Graeme Philipson is a director of Strategic Publishing Group and Strategic Research. He has served in senior research and consultant roles for the GartnerGroup, the Butler Group, IDC, Compass Research, and the Yankee Group Australia. He also is a former editor of Computerworld Australia.

Later this month IBM Australia will hold a seminar to promote its venerable AS/400 computer which has just undergone yet another major upgrade. I have been asked to speak at this event because I have been following this machine closely since its release in 1988, 12 years ago this week.

During that time I have become a big fan of the AS/400. I believe it is close to being the world's most underrated computer. It is virtually ignored by many consultants and journalists, yet it is the most widely used non-PC computer in history.

I became interested in the AS/400 many years ago when I noted a major disparity between its usage and its profile. More than 500,000 AS/400s are in use around the world, including more than 3,000 in Australia. Nearly 20 per cent of organisations that use computers to run their businesses use the AS/400. If IBM's AS/400 division were a separate company, it would be larger than any other computer company except IBM.

Yet if your entire knowledge of the computer industry was confined to what you read in the press, you would think the AS/400 did not exist, or was a fading machine used only by a small minority of users.

Most commentators and journalists conveniently or ignorantly overlook the facts about the AS/400, because the AS/400 is not sexy. The "AS" stands for "Applications System", which goes a long way towards describing the sort of machine it is, as well as the reasons for its reputation.

When IBM designed the AS/400 and its predecessors the System/34, System/36 and System/38 it designed it as a computer for running commercial applications. It did not have a lot of bells and whistles and it was not designed around a particular chip or operating system.

At a time when most of the industry was designing computers to abstract technical specifications, the AS/400 was designed to handle real world applications. It uses, for example, a unique memory management system that treats disks and RAM as one large memory space, anathema to the purists but which makes it very easy to program and operate.

The AS/400's massive success in its target markets (especially in areas like manufacturing and distribution, where it is undisputed king) indicates that IBM has been doing something right. That something has been talking to the users of the technology in a language they understand, which essentially means "solutions", an overused term in the computer industry, but a relevant one.

People don't buy computers to have a computer, they buy a computer to do word processing or get the general ledger out or to build a better mousetrap. In an industry dominated by technology, the computer press and the computer consultants and the computer vendors all too often concentrate on the so-called "speeds and feeds" of the technology, and forget that the technology exists only to help people do their jobs more efficiently.

More than any of its competitors, the AS/400 has been a computer that has allowed people to get on with the job. Most people simply don't care about operating systems or chip architectures. They just want to do a job, and they will buy what they perceive to be the best machine to do it.

The AS/400 has been unlucky in that its popularity has been overshadowed by the rise of Unix and so-called "open systems" and, more latterly, machines running Microsoft's Windows NT, now renamed Windows 2000. These architectures have succeeded for very different reasons than has the AS/400.

Unix was an off-the-shelf operating system that anyone could use, at a time when a lot of computer manufacturers were looking for just such an operating system. Many of them were newer and smaller companies, and they lacked the resources to build their own operating systems.

They were able to take advantage of massive improvements in microprocessor and memory chip technology to build very powerful machines at very low prices, and many of them succeeded because of those advantages. A whole swag of software companies also rode that wave, providing databases and other systems to make the new Unix hot boxes suitable for commercial applications. Unix itself, although cumbersome, was very flexible, and gradually the Unix market matured to the extent that it was capable of performing the sorts of turnkey applications that the AS/400 was so good at.

Similarly, Windows NT will run on virtually any Intel processor, and so has helped commoditise hardware.

In the early '90s IBM realised that it had to do something about the Unix challenge. It released its own Unix machine, but it also significantly upgraded the AS/400. At the same time that Unix was growing in use, a new phenomenon called client/server computing had become the big talking point.

Client/server was the most interesting new technology of its era. Stripped of all the hype, the term simply describes getting PCs and central computers working well together, sharing data, processing it efficiently, and helping the users of the technology do their job. Unix was good at this, and the AS/400 was not.

Client/server was a sort of code word describing the many changes in the computer industry in the first half of the 1990s the rise of the graphical user interface and relational databases, the empowerment of end users, the incorporation of corporate data into end-user applications, and the like.

In 1995, IBM announced changes to the AS/400. It rewrote and opened up the OS/400 operating system. It redesigned the hardware to use its PowerPC chips (the same ones as used in IBM's RS/6000 Unix machine). It introduced a number of client/server and inter-operability features.

These improvements have continued with the latest releases, which include the first commercial computers in the world running IBM's new Silicon on Insulator (SoI) chip technology.

Yet most people still think of the AS/400 as an old-style computer. They probably always will. IBM will probably never completely overcome the AS/400's image problem, but the good news is that hundreds of thousands of happy AS/400 users simply don't care.