Happy 18th Birthday, AS/400; Time to Leave the Nest
June 19, 2006 Timothy Prickett Morgan
In some ways, it is a pity that turning 18 years of age does not, in some ways, give the Application System/400, which is now called the System i, some rights of its own that transcend that of its parent, IBM. When we turn 18 in the United States, we can vote and we can join the military, but we still can’t drink legally, even though a lot of us certainly drink against the law. Maybe it is time for the System i to stand up for itself and, in a very real way, against the way that IT is conducted in the data centers and departments of the world. And maybe in few years, we can have a celebratory drink.
I don’t know about you, but I am utterly amazed that I am still writing The Four Hundred after all these years. Shortly after I graduated from college, I took a job with a boutique technical publisher in the SoHo district of Manhattan to become the editor of a newsletter dedicated to the AS/400 platform; I knew very little about computers back then (but more than most people), and I have learned a lot since then. I’d like to thank you all, once again, for the education. Anyway, this newsletter has changed names once, and then changed back again, and I have worked for four different publishers (two of them owned by myself) in all those years, but I am still writing about this wonderfully eccentric and unique server platform.
And I still get yelled at for calling it the AS/400, as I did once again by a staunch System i supporter last week. That’s not just habit. It is also defiance. I want to remind everyone that this machine used to be about applications, and the battalions of independent software vendors who used to code explicitly for the Application System/400. I want to remind IBM that it is the army customers who love their homegrown and, sometimes, third-party RPG applications who make this business work, and that if IBM needs to get back to basics, it needs to make RPG a true, Web-enabled, SOA-perfected, wonderfully easy programming language. Among the many very smart things about the AS/400–virtualized hardware, integrated relational database management system, integrated set of development tools, programming languages with native and fast access to database information–is the fact that RPG was a very good language in which to implement business rules and create applications to run businesses. The 5250 protocol was very streamlined, and used very little resources. Applications could run–and run well–on distributed machines with fairly modest processors and co-processors. This is what made the AS/400 such a profit engine. It was, in essence, a legion of Macintoshes trained to behave like a mainframe.
As the AS/400 turns 18, as it will on June 21 this week, IBM is leaning pretty heavily on the “i” word–or rather, just the “i”–as its marketing message, and very soon the company is expected to ramp up a bunch of advertisements that hammer out this message of integration and infrastructure simplification. I am skeptical of this approach to promotion. When all of IBM’s platforms can run multiple operating systems–all of them run Linux and Unix, and two of them run i5/OS (formerly known as OS/400), and one runs z/OS (formerly known as MVS and OS/390). Only one platform runs Windows–that would be System x (formerly the xSeries), and no, I don’t count an Integrated xSeries Server co-processor, an Integrated xSeries Adapter linking to another outboard xSeries machine, or a Windows box linked to the iSeries by an iSCSI port as a native implementation of Windows. So I personally don’t think the i for integration means as much as some people at IBM believe. Big Blue itself has diluted the value of this integration theme by essentially making its other server lines look like the AS/400 and then the iSeries. I didn’t do that; IBM did. The decision in 1995 to merge the AS/400 and RS/6000 hardware lines was the first step; the Mach Project from 2000, where Linux and logical partitioning was spread across IBM’s lines and eServer was born was the next step; and Project Polaris, from last summer and announced as the IBM Systems Agenda, took yet other steps to remove even more of the distinctions between IBM’s servers, which are now called systems in a throwback to 1960s and 1970s naming conventions at Big Blue.
I really wish that IBM had just called the machine the System/400–if it had to use system by itself–and the operating system OS/400 and just gone all the way back to the future. This was the right thing to do. To admit that the whole i, p, x, z differentiators were stupid, or at least have outlived their usefulness. You cannot solve your problems by changing your name. But you can use the name people know you by, and then try to solve your problems.
If the AS/400 were a young adult, it would be time for it to figure out what it was going to do with its life. And frankly, I think the OS/400 community has a better sense of who this 18 year old is, what it needs, and where it wants to go than IBM does. In a just world, the OS/400 community would get the backing of some venture capitalists and liberate the AS/400 from Big Blue, since it never has quite known what to do with it. We’d still go to IBM and buy System p servers on which to run OS/400, but we would determine what OS/400, RPG, and 5250 were and what it costs. Of course, I would bet Sam Palmisano’s last dollar that if the OS/400 community could do a hostile takeover of the System i business, the first thing that community would do is create projects to port the platform to X64 and other RISC architectures, thus diminishing the need for expensive Power-based hardware in the first place. (Psst. Don’t tell IBM that.)
As much as I love hardware, and as much as I hate to admit this, no one ever wanted an AS/400. They wanted OS/400. This is all about the software, and somewhere along the lines, IBM really forgot that. Going broke, as IBM nearly did in the early 1990s, will mess with a company’s collective mind, and moving to a services-oriented business plan was the exact and precise antithesis of all of the value propositions of the AS/400 line.
Even with System i sales as low as they are–probably around $2 billion in 2006, including systems, storage, and software–it would still be very expensive to acquire this piece of IBM, since I think, even today, more than half of that money falls to the bottom line. IBM only had net income of $12.2 billion in all of 2005 against $91 billion in sales. The System i business might only be around 2 percent of Big Blue’s sales, but it is probably around 10 percent of its profits. This is not the kind of business you let go of, even if your hugging it is doing it harm.
So, what IBM should probably do is act like it has let go of the System i business, even if it doesn’t actually let go. If the OS/400 community is right–and I think we are–then spin out this entire business as a wholly owned subsidiary. Make Frank Soltis its chief technology officer and make Mark Shearer its chairman of the board and chief executive officer. Let IBM be a contract manufacturer for hardware, and if the Application Systems Company can find better or cheaper hardware and make it up with software profits–like Microsoft has done so brilliantly, so don’t even try to tell me this doesn’t work–then take that risk. Give this new company some resources to see if it can grow, but get the profit monkey off its back so it can prove that, in the long run, the AS/400 value proposition is the right one for the 21st century.
Because, Big Sam, I can tell you two things. First, the IBM Global Services vision of the world from the 20th century is not what the future holds, and you are going to learn that the hard way unless you change your ways. You can move all of IBM to Bangalore, Beijing, and Bangkok, and that still will not change what I am saying. Second, sooner or later, the open source community is going to figure out how to build an open source, integrated, hardware, software, and application stack–and then IBM’s entire business is at risk. You think the 1990s were bad? Forgeedaboudit, as we say in Brooklyn.
It would be nice to lead that army instead of being overrun by it. So get out of the way and let the OS/400 community lead.