The Persistence Of The IBM i Platform
February 5, 2018 Timothy Prickett Morgan
Talking about switching platforms is one thing, and it is relatively easy if you are talking about moving from Unix infrastructure servers based on RISC servers to machines that are going to run essentially the same workloads on Linux systems based on X86 processors. Leaving the IBM i fold, particularly for customers who have created their own applications, is another matter entirely.
That, in a nutshell, is one of the main reasons why the IBM i platform, in its many different incarnations in the past 30 years, has persisted. The database is the stickiest piece of software in the datacenter, and the applications that touch this database, encapsulating and embodying the transactions that literally define the business, are the next stickiest things.
In survey after survey done by HelpSystems in the past four years, IBM i shops were asked if they planned to move all or some of their applications off the IBM i stack, and went even further and asked what platforms they would move those applications to. The number had been trending down in the past three years, but in the 2018 IBM i Marketplace Survey, the percent of customers surveyed actually went up. What gives?
Back in the 2015 survey, 11.1 percent of the customers polled said they were planning to migrate all of their applications to a new platform. This seemed like a pretty high number. It fell to 8.5 percent of those polled the following year, and then dropped to 7.3 percent in the 2017 survey. All of these numbers, quite frankly, seem a bit high to me, based on the actual bleed rate that has been seen in the actual base over time. And a few percent of bleed rate, by the way, can have a big impact on the base over time, so I am not trying to say this is not important, or that there does not continue to be competitive and management pressure on all platforms to move to Linux or Windows Server. That pressure is always there, and has not abated one bit.
But the fact that in the 2018 survey, 9.8 percent of those polled said they were planning to migrate off the IBM i platform is not necessarily a cause for freaking out. As far as I can tell (mostly from anecdotal evidence through talking to resellers and customers) we did not actually see an exodus in 2015, or in the years that followed. The best laid plans of mice and migrations often go awry.
We are not seeing a massive shift from homegrown to third party applications, as most certainly did happen in the late 1990s and early 2000s, and the current 2018 survey shows a very robust 70.1 percent of customers are using homegrown applications; the percentages also suggest that a large number of shops are using third party software. (Our guess is that some companies write their own apps, some do half and half, and some go all in with a software supplier.) If we see that homegrown application figure go down – it actually went up a few points from 2017, when the question was first asked in what I think of as the correct way, with homegrown software being an option – then that will probably signal a problem. But the cost and disruption of shifting platforms while also trying to transform and modernize applications is something that companies might talk about doing, or think about doing, but there has to be a compelling economic and technical reason to make the jump. If one company buys another and they want only one set of applications, IBM i can definitely lose out footprints. But this is a rare situation.
Having said all of this, we will be watching the 2019 survey, in the wake of the Power9 ramp, to see if there is more or less talk about migration. That, we think, will be a stronger signal of a real shift, if there is one.