Evangelizing Mainframe
Print Email

Saving Money With Mainframes

As we head inexorably toward the end of the current financial year, we still find ourselves living through recessionary times—whether governments choose to call it such or not. The consequence is that many mainframe sites are being asked to do more with less. And some are being asked to justify the cost of their very existence. But at all mainframe sites, people are looking for ways to save money.

The trouble with financial years is that they focus on just that—a year, 365 days—whereas the costs associated with a mainframe are amortized across a much longer timeframe. This can sometimes make it harder to have a realistic conversation with the CFO about overall costs. For example, a survey carried out a few years ago for the Virtual IMS user group found that many IMS sites were running older versions of the software. They didn’t feel the need to go the latest release in order to use the latest facilities; everything worked the way they wanted. An obvious consequence of this is that there was no expenditure on updating applications, and the software costs for IMS were pretty stable.

Mainframes tend to run almost continually, and there can be very long periods between reboots—especially unscheduled reboots. So, there’s generally little cost savings to be had from running the processor 24/7—this is what typically happens already. Some sites have looked for savings with specialty engines like System z Integration Information Processor (zIIP), System z Application Assist Processor (zAAP), and IFL. These process program instructions rather than using the general purpose processor (GPP), which is also used to calculate billing—million service unit (MSU) rating. Of course, there is a cost involved in turning on a specialty engine. But sites can probably make savings overall because paying for a specialty processor (and the cost is quite high) can remove—or certainly postpone—the need to spend even more money on a much higher-ticket item—a new mainframe. Again, this saves money over a longer timeframe than a year.

Another way to save money is look at the software you’re running and see whether there are equivalent but cheaper alternatives. Many ISV Web sites make predictions about massive savings that organizations can attain by switching to their software. Even if the product or suite of products is cheaper, there are always the additional costs of training and down time while the software is installed. Apart from researching thoroughly to confirm that the cheaper software does exactly what you want it to, it’s important to consider any additional costs. But once all the other costs have been taken into consideration, there might well be savings to be made by changing software vendors.

But almost every purchase, these days, seems to benefit from starting with a strong negotiating position. It’s not like going to the supermarket where everything has a price stamped on it—and that’s the amount you pay. Nowadays it’s more like a souk, with much “too-ing and fro-ing” before a final price is agreed upon. Customers can make savings by negotiating good deals.

Staffing is an area where savings have traditionally been made, but, nowadays, mainframe departments are already reduced to very few staff. Sometimes looking at the bigger picture can help. It can be useful to integrate non-mainframe and mainframe IT staff. Provided suitable training and documentation is available (and I know there’s a cost involved), it can lead to fewer IT staff overall being required. I’m not suggesting that the SharePoint expert should necessarily start fiddling around with CICS, but it makes sense for the distributed Linux guru to learn how Linux on System z works. And if you’re a DBA, your experience may make learning DB2 faster, easier and fairly straightforward. In fact, moving the whole distributed Linux system onto a mainframe would undoubtedly lead to huge savings for the IT department—enterprise-wide savings.

Lastly, storage costs can be lower on distributed systems than on mainframes. It’s worth investigating whether there is a business case for storing archive data, which may be needed for legal reasons to be available for many years, off the mainframe, and keeping the DASD full of live data.

Unfortunately, there are still many managers that see migrating applications off the mainframe and onto distributed systems as the best cost-saving method. In the long- and short-term, these savings can be hard to validate. There are too many “extra” costs that need to be taken into consideration.

Whatever strategy you adopt, good luck with your year-end meetings.

Trevor Eddolls is CEO at iTech-Ed Ltd., an IT consultancy. For many years, he was the editorial director for Xephon’s Update publications and is now contributing editor to the Arcati Mainframe Yearbook. Eddolls has written three specialist IT books, and has had numerous technical articles published. He currently chairs the Virtual IMS and Virtual CICS user groups.

Posted: 3/20/2012 4:00:07 AM by Trevor Eddolls

Print Email

Join Now!