Big Blue Gives A Solid Installed Base Number For IBM i

August 6, 2012 Alex Woodie

Estimating the size of the IBM i installed base has traditionally been one of the more fun and entertaining exercises occupying the time of analysts and press covering the IBM midrange marketplace. Where else can one combine math, astrology, and tea-leaf reading skills to generate a number that everybody in the business seems to want? With a single disclosure in the latest IBM i roadmap, Power Systems general manager Collin Parris has all but killed demand for that particular skill set, at least for the time being.

In the latest IBM i roadmap, which was published on July 11, Parris gives us the number we have all been asking for: 150,000. According to Parris, the IBM i server (not multi-OS Power Systems server, but what we all used to call an AS/400, iSeries, and System i) is used by “over” 150,000 companies around the world. Undoubtedly, not all of these customers (which aren’t only businesses but are also non-profit organizations and governmental entities) are running IBM i on the latest Power Systems gear, but we can infer that they all currently run some iteration of what we now know as the IBM i family. That is, they are IBM i shops.

Parris disclosed the number in his introduction to the latest IBM i roadmap document, which was published on July 11 and which bears the name IBM i: An executive guide to IBM’s strategy and roadmap for its integrated operating environment for Power Systems. You can download the paper at www-03.ibm.com/systems/power/software/i/strategy.html.

“IBM i is the integrated operating environment for IBM Power Systems, which is used by over 150,000 businesses around the world,” Parris writes. While this statement is a bit vague (Power Systems, of course, can refer to Linux and AIX environments as well as IBM i environments), the specificity of that magical number was bolstered a few pages later in the paper when IBM stated: “IBM i is used by over 150,000 companies in more than 115 countries around the world to run their business applications.”

This number is significantly better than the tea-leaves have been indicating. In a 2010 blog entry, IBM i chief architect Steve Will wrote that there were “easily over 100,000” enterprises running IBM i. In enterprise speak, this would have indicated that the installed base might be in the 110,00-to-125,000 customer range, but certainly not 150,000 or more. If the number were considerably larger than 150,000, the enterprise-speak tendency to would encourage talking of the installed base as “nearly 200,000.” Thanks to Parris, we have some clarity into the actual number, and the actual number is actually a good one.

So why is this so important? Why do people care so much about knowing, with a degree of certainty that only IBM can provide–and no, IBM, we never believed your excuses that “it’s hard to quantify the installed base” since you make the best quantifying machines the world has ever seen–exactly how large the IBM i community is? The short answers to these questions parallel basic human emotions. IBM i people care about how big or how small the IBM i marketplace is out of basic fear, anger, sadness, love, and joy.

IBM i professionals are underdogs, and must constantly defend and justify and fight for their choice of business machine and profession. The IBM i server is a weird but adorable entity that defies explanation. This life of uncertainty lends itself to fear. If the IBM i installed base number had come in at less than 100,000, there would have been a considerable fear response in the IBM i world, because there is constant fear that the IBM i marketplace will disintegrate and fall apart faster than it has.

Make no mistake about it: the IBM i market is getting smaller. While Parris’ “over” 150,000 number is considerably better than Will’s “easily over 100,000” number, the difference is almost certainly due to the lack of clarity and transparency that IBM executives historically have given to the question. If the installed base had grown from 125,000 to 150,000 over the last two years, IBM would have shouted it from the rooftops (at least we hope it would have).

Since real growth can’t legitimately be expected at this stage in the IBM i machine’s life, the only question that matters is: at what rate will the installed base and the overall market continue to shrink? With “over” 150,000 distinct customers (actually “more than” 150,000, as good grammar and math would indicate), the IBM i installed base is at the same point that it was in the early 1990s. The AS/400 was a huge hit when it was launched in 1988, and grew quickly as mainframe shops migrated to the new midrange machines.

But those early AS/400s were much more expensive back then, and IBM was raking in nearly $4.8 billion per year in revenues from the Rochester creation. By the end of the decade, the AS/400 installed base hit its peak of about 275,000 customers, although it was bringing in a tad less, maybe $4.7 billion a year, according to the estimates of IT Jungle‘s chief tea-leaf reader and IT astrologist Timothy Prickett Morgan.

By TPM’s estimates, the installed base was about 250,000 in 2002, and about 220,000 in 2005. Since then, according to Parris’ disclosure, the installed base has shrunk by an average of about 11,500 customers per year over the last six years. The good news is that’s considerably better than previous loss estimates, which pegged the annual migration from the main IBM i herd at about 22,000.

So what’s the takeaway from all this? If Parris’ disclosure is accurate–and there’s no good reason to believe otherwise–then we can rest a little easier knowing the IBM i business is not shrinking as fast as some had feared. While it’s nowhere near the business it was 20 years ago, the specter of mass defections from IBM i to the worlds of Windows, Linux, and Unix have been scarier in our minds than in reality.

Less than half of the customers that have ever been an IBM i shop have ditched the machine for something else. The bulk of the IBM i pack is still intact, even with ferocious forces like cheap X64 servers and a worldwide economic collapse, continually chipping away at the edges. The end of the world of IBM i will come, eventually, as it must for everything. But it’s not in the foreseeable future, at least according to today’s tea leaves.