Forced Windows Migration Failures
November 28, 2016 Dan Burger
Here’s what 90 percent of the companies don’t do before, during, and after making the decision to migrate from IBM i to Windows. They don’t plan the way they would for any other major business event. Quite possibly that’s because management doesn’t understand the system they currently have and how it works, nor do they understand what the new system entails. Hasty decisions can be costly decisions. Hundreds of thousands of dollars are wasted.
A recent blog post by Bob Losey on LinkedIn casts a critical look at these migrations and why so many of them are failures.
In a telephone interview with IT Jungle last week, Losey expanded on the blog topic, which attracted more than 1,000 readers on LinkedIn plus an unknown number from a post on the social media site Google+.
His perspectives on the migration topic come from talking with IBM midrange users for more than 20 years. Losey has been working in the IBM i market since 1979, when he founded his company Source Data Products. He started hearing rumblings about companies leaving the AS/400 platform for a variety of reasons in the mid-1990s.
Losey has a list of reasons why IBM i ERP migrations crash on the rocks. It doesn’t matter if the ERP is from a software vendor or if it’s home grown. The decision to migrate is largely based superficial thinking and is wildly optimistic in terms of how quickly and inexpensively it can be accomplished.
“At a management level, the complexity is not understood,” he says. “There’s no consideration for figuring out how all the subsystems fit together.”
He backs that up by saying managers and power users have not taken the time to understand their applications in depth. They don’t understand the transactions. As they look at alternative systems, they are not paying attention to fields and files. So, they may not get a good conversion of data from one system to another.
That’s just for starters.
In many instances, there are files that are redundant. Temporary files are often updating permanent files and migrations may not account for where the permanent data is located. So, in the process of mapping data from one system to another, data is sometimes pulled from the wrong files.
Losey says when focus is on the GUI and the green screen, without attention to the customized functionality in the existing system, he’s seen instances where subsystems can’t be ported and so the migration from the IBM i platform can’t be completed. “The subsystems are too expensive to write from scratch and the new system doesn’t convert those programs. That can go on for a decade,” he says of a problem that often surprises the migration proponents. “There’s a lack of planning, lack of integration, and lack of business understanding that all contribute to lengthy implementation efforts and cost overruns.”
Losey’s blog indicates that companies will spend five to six times more on software, hardware, infrastructure, and staffing when implementing a Windows migration compared to remaining on the IBM i platform.
That’s pretty hard to pin down because there are a lot of factors to take into consideration, including the cost of doing nothing or the cost of upgrading an existing system to reach the expectations of the end users. However, Losey’s point that it takes more people to manage the Windows environment is valid. Scale out systems require more people than scale up systems and system automation features still favor the IBM i platform, but that gap has narrowed. The flip side of that is that IBM i management skills are difficult to find and likely to be more expensive when found.
IBM i skills are being lost to retirement, he concedes, but it’s not a doomsday scenario.
“There are still a lot of good people in their 30s and 40s working on i platforms,” he says. “I think management sometimes doesn’t know where to look.”
To make more points regarding an unattainable return on investment on Windows migrations, Losey sites a disregard during the decision process for factors such as the necessity for Windows servers to breakdown ERP apps with individual servers for such things as accounts payable, inventory, and purchasing, which add server farm complexity and more complex client-server networks.
At the conclusion of Losey’s blog, he cites a consultant friend who suggests a Windows migration preventive program begins with educating management about the problems that can be avoided and the money that can be saved by remaining on the system.
Some biases that favor Windows will never be overcome and, although Losey doesn’t promote it, there are some migrations that make sense for a business. But a deeper understanding of what a migration actually entails with a more accurate time frame and cost estimate could derail some knee-jerk decisions that are potential disasters.
You can read Losey’s “Where Is The ROI To Leave IBM i In Favor Of Another Platform?” blog here.