News Stay informed about the latest enterprise technology news and product updates.

Moving apps off mainframe may have unintended results

Migrating apps off the mainframe may result in unintended negative consequences, as illustrated by this hypothetical account of a capacity planning meeting.

Migrating applications off the mainframe may seem like a good idea, with CPU consumption and cost savings benefits. But if the application's data model isn't fully understood, these migrations can become costly and may require your organization to invest in additional computing resources. In this column, mainframe expert Robert Crawford uses a hypothetical capacity planning meeting to demonstrate the unintended negative consequences that can come from migrating apps off the mainframe.

The conference room is dim, lit only by the screen showing a presentation title page. Around the table, men and women are talking quietly about the weekend or swapping jokes.

Normally, these people would be thoughtfully deliberating over CPU and resource consumption graphs on the screen, but today, the Rock Star Architect (RSA) is supposed to attend and he is late. The RSA is always late because he is busy doing Very Important Things. Doing these things, he's rarely able to stay in one place for more than half an hour, let alone attend low-level meetings like this one. And so they wait.

About five minutes later, the RSA breezes into the room. The senior manager greets him as he takes his place. The rest of the people in the room look expectantly as the RSA settles into his chair with a faint smile. Now everything is in readiness.

The attending senior manager flips through the first few slides in the presentation while the RSA looks on thoughtfully. The manager gives an overview of the capacity planning team's methodology and data sources. He expresses their hope that careful study of resource utilization and application tuning will save money and postpone equipment acquisition. To this point, the presentation is full of colorful, oversimplified and unambiguous graphs to prove how well things are going.

Now one of the analysts in the room takes over the presentation, as it is time to move on to more technically detailed information. The other analysts perk up, as this part of the presentation contains data they themselves collected and analyzed. They have also been working very hard to influence the graphs, almost willing them to bend into the intended shape. The RSA also likes this part of the presentation. He will be sure to ask several well-placed questions to show he still has a technical grasp of the essentials despite his lofty position.

The next chart shows mainframe CPU consumption. The graph displays a discernable downward trend with a dotted line trend for the future. The analyst running the presentation explains that the reduced consumption is the result of migrating certain applications off the mainframe onto other platforms. She reels off some figures, such as the percentage of CPU, chargeback dollars saved and deferred upgrades.

On cue, the RSA smiles benignly. His long, hard work is now being rewarded with the affirmation of his analysis and predictions. He compliments the analyst making the presentation. To the others he notes how much this small effort will save in expensive mainframe costs. He adds how this represents the first step in an ongoing modernization of the enterprise's IT infrastructure, with a new paradigm that will synergize their applications, making the enterprise more nimble and able to compete in the 21st century. Inwardly, the RSA thinks about how he's going to present this information to executive management.

For the next segment in the presentation, a man replaces the analyst in the chair. After a few introductory remarks he goes to the next page, which contains a graph of midrange platform utilization. Unlike the previous graph, this one shows a decided uptick toward the end. The predictive dotted line shows a further increase with a slightly higher slope.

The RSA frowns and asks the analyst for the cause of the increase. The analyst explains that the growth in CPU seems to be associated with an application recently migrated to the midrange platform.

The RSA smiles again. He asks if the recently migrated application has been tuned yet.

The analyst relates how the application team has carefully gone over the code for the past few weeks. The team found a few bad database calls and some unnecessary loops, but the easy things have been done. The analyst adds that the application's data model wasn't well understood during the migration and additional SAN storage and database servers will be required shortly.

The RSA frowns again and is about to say something else when the hardware provisioning manager interrupts to ask how soon the new servers will be needed.

The analyst notes that this is a rush order and the paperwork should be complete by the end of the week. The hardware provisioning manager scratches his head and wonders aloud if he can get space and cabling lined up within that time.

This starts a lively discussion among the analysts. Some of them are trying to figure out what the projection means. Others are talking about pending projects that are supposed to migrate more work to the midrange servers and may drive the utilization curve higher.

The RSA grimaces spitefully at the disobedient graph. Eventually the side conversations wind down and a few people look expectantly at him for some wisdom or advice. The silence breaks when the presenting analyst adds that he needs the Rock Star Architect's help to get the new hardware despite the current procurement freeze.

The RSA's eyes widen a little, then his Blackberry rings. He quickly grabs it out of its holster and mumbles a few words. He gets up and makes his apologies for leaving the meeting early, but there are some Very Import Things he has to attend to that cannot wait. He glides out the door.

The room seems a little emptier now without the RSA. The remaining people quietly discuss what they have to do in the next few months to deal with the coming workloads.

Soon after, the meeting breaks up.

ABOUT THE AUTHOR: For 24 years, Robert Crawford has worked off and on as a CICS systems programmer. He is experienced in debugging and tuning applications and has written in COBOL, Assembler and C++ using VSAM, DLI and DB2.

What did you think of this feature? Write to's Matt Stansberry about your data center concerns at

Dig Deeper on IBM system z and mainframe systems

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.