IBM i Bucking The Trends, Year After Year
March 15, 2021 Timothy Prickett Morgan
There are all kinds of stability that we consider when we choose and use systems. In the IBM i market in particular, we often talk about stability in the sense of the good programming practices that companies or their third-party software vendors have for the applications that they run. Or we might talk about the underlying stability of the operating system, relational database, or middleware software on which these applications depend. Or digging down further, we might talk about the reliability and longevity and predictability of the Power Systems hardware that underlies it all. And then, of course, there is Big Blue’s own commitment to the System/38, the AS/400, and their follow-ons for more than four decades.
But there is another kind of stability that is important to the IBM i community, and that is the constancy of the base itself. This is just as important as all of the other aspects of stability we mentioned above: The stability of the customer base and its intention to continue to run the applications at the heart of their companies atop the IBM i platform. Here are some datapoints that express this constancy of intent, which comes from the 2019 IBM i Marketplace Survey done by HelpSystems, which has four years of survey data. Take a look:
And here is a similar but slightly more sophisticated set of questions about the commitment to the platform that comes from 2021 IBM i Marketplace Survey:
Remember that the survey report comes out in the winter (January) of the year following the fall of the year (October) when the survey questions are asked. So 2019 data in the chart is actually from the end of calendar 2018, and thus the so-called 2021 data shown in the chart immediately above is from the end of 2020, when we were in the middle of the pandemic.
No matter who answers the questions, no matter what year we are talking about, the interesting thing is that the intentions of customers do not seem to change very much. And really, that is an amazing set of statistics shown above. Somewhere north of 40 percent of the customer base always says they have no plans to change their investment in the IBM i platform, year in and year out. Somewhere between a fifth and a quarter of the base tends to say they will be increasing their investments in the platform. So basically two-third of the base is steady or up. Somewhere between 8 percent and 10 percent of the base is always threatening to move all of their applications to a new platform, and a somewhat larger percentage say that they are going to migrate some of their applications to another platform. And about an eighth of the base tends to say they are not sure.
Here is the funny bit. If people moved off the platform as expected in the 2015 survey data shown above, the installed base would have taken a tremendous hit if applied across the entire base and assuming that this survey data was representative of the entire base. And as far as we know, there is no indication that this happened. And that is a beautiful thing.
What we don’t know is if this kind of stability is typical in the IT sector, but we have some anecdotal hunches. There are some platforms that, so long as their creators keep them on modern hardware and keep the systems absorbing new technologies as they become available and go mainstream. Windows Server has been commercial grade for more than 25 years, and Linux for about 20 years. The pace of change for these platforms has settled down and a lot of the innovation in these platforms is being done way up higher in the data storage and data analytics portions of the stack. AIX on Power platforms (1990) is two years younger than the OS/400 and its AS/400 hardware (1988) and we don’t think it has fared better at maintaining its base than IBM i has; in fact, as percent shrinkage in unique customer and machine footprints, AIX might have done quite a bit worse than IBM i. We know for sure that Sparc/Solaris and PA-RISC/HP-UX from Sun Microsystems/Oracle and Hewlett Packard Enterprise have done far worse in as much as they are absolutely defunct and date from around the time of the AS/400 in the enterprise. Our point is, we don’t think IBM i is any worse, and probably better off in these regards, than its peers in the small, medium, and occasionally large datacenters of the world.
It is important to be reminded of that sometimes.