A Platform Of A Certain Age And Respectability
June 20, 2018 Timothy Prickett Morgan
Outliving your enemies is one of the best kinds of revenge, if you happen to have any enemies. But the people of Rochester, Minnesota, did not really have many enemies, although they did build a system that has outlived its many rivals and that continues to co-exist alongside newer platforms that are, quite frankly, still difficult and expensive to use.
That the AS/400 is celebrating its 30th birthday this week, and the IBM i platform running on Power Systems is still around and still a viable business, is nothing short of remarkable. We know this because we just remarked upon it.
But seriously. The many rivals of the OS/400 platform and its follow-ons since that June 21, 1988, launch of the Application System/400 are now gone or not even on life support. We can all rattle them off, but the important ones that drove innovation for OS/400 and its children through to the current IBM i are DEC’s VMS for the VAX and Alpha systems, Hewlett Packard Enterprise’s MPE for the HP 3000 and HP-UX for the HP 9000s, and Sun Microsystems’ Solaris for the Sparc systems. You could throw in SCO Unix, Novell NetWare, and a slew of proprietary operating systems in Europe and Japan, and while you are at it, you should probably also include the IBM System/38’s CPF operating system and the IBM System/36’s SSP operating system. Even OS/2 and its PS/2 platform actually predate the AS/400 by 10 months – and they are long, long gone.
And among all of these and other platforms, IBM i is still out there – probably at somewhere around 125,000 unique customers around the global and maybe on somewhere between 250,000 and 300,000 systems – doing useful work. That is less than half the peak number of customers set back in 1998, mind you, a decade after the AS/400 was launched and the base reached 275,000 unique customers and probably somewhere around 700,000 to 800,000 systems. But that was a very different world, and one where the systems market was a lot smaller than it is today in terms of money and certainly in terms of shipments.
Many more things in business are computerized and automated today than was the case three decades ago, and these operations are sophisticated enough that it takes a lot more computing oomph to get the jobs done. But here’s the thing: Good old fashioned online transaction processing is something that every business has to do, and even a hyperscaler like Google uses traditional applications to keep the books and run the payroll. (You didn’t actually think Google runs its back-end on its own cloud, did you? Or that it writes its own enterprise resource planning software? It is very likely that Google is an SAP or Oracle shop, like we know IBM is.)
We know that the IBM i platform is more than an OLTP machine, and that it has evolved within the constantly changing environment of modern datacenters in the past three decades. This is a testament to the ingenuity and continuing investment by Big Blue in its Power chips, its Power Systems servers, and the IBM i and AIX operating systems and their add-ons. Linux came along two decades ago and has helped bolster the Power platforms, but not to the same extent that Linux has helped the mainframe stay relevant. The mainframe had much higher costs for hardware and systems software, and therefore lower priced Linux engines on mainframes exhibited a kind of elasticity of demand that IBM wishes it could get for IBM i and z/OS, the main operating system on its System z mainframes that has been around in one form or another for more than five decades.
Unlike many of the people who worked at IBM on the AS/400 platform before and after its launch, we came to it after the launch and saw an opportunity to provide strategic and tactical advice to customers who, by and large, were going to be coming from a simpler System/36 world and were going to be facing some of the complexities of online transaction processing setups that looked and smelled a whole lot more like the System/38. We saw a disconnect, and an opportunity. We also saw a well-engineered platform that would have longevity and that we could base a business upon – just like those AS/400 customers the world over. While we never used the system – except as a web server for a few years – we were absolutely as dependent upon that system as any user could be. The Four Hundred is absolutely tied to the ups and downs of the AS/400 and its progeny, and it has always been our pleasure to serve this friendly community. There is not another machine like it in the world, or another community like it that supports it. Perhaps there never will be again, because this takes a certain set of initial conditions, starting with an IBM division that was cast out to the hinterlands of Rochester, Minnesota, and some software drawn in from Toronto and Atlanta. The conditions are unique, and so is the system and therefore so are the users.
So here is to hoping IBM i is around for a long time to come. . . . It has outlived the expectations of many, after all. As long as people do business, as long as there is online transaction processing and midrange and large enterprises, there is every chance that if you look over a counter, in a data closet, or in a full-blown datacenter, you will find the child of an AS/400 running there.