Why is a software upgrade that instantly saves you 5 to 10 percent delayed while you’re looking to reduce costs almost everywhere in your company? I’m talking about the fact that 75 percent of IBM DB2 customers haven’t yet taken advantage of the superior resource efficiency in DB2 10 for z/OS. To better comprehend why we allow this type of waste in our IT departments, we need to understand how we ended up where we are today.
How We Got Here
For many years, the expression “the cost of doing nothing” has been used by authors to explain why accepting a status quo isn’t only the wrong thing to do, but it’s dangerous and can put your company at risk. When noted technology and business author, Nicholas Carr, published “Does IT Matter?” in the May 2003 edition of the Harvard Business Review, it created quite a stir. He proposed that the added value of IT for the business was questionable at best.
Intense discussions followed, causing many IT professionals to look at things more from a business perspective, and at the same time, made them realize things needed to change. But in 2007, soon after Carr’s book of the same title became popular, the first economic cracks started to appear. Armed with the arguments from the book, many C-level managers decided that IT could (and should) spend a lot less money. Finally, a hammer that made everything look like a nail; here was an easy way to save money and reduce future budgets. After all, they ran a financial/retail/oil business, not an IT business!
In the years that followed, the staff who managed the mainframe was particularly hard hit—slowly reduced to the numbers we see today. In fact, some customers now run part of their IT Infrastructure with 50 percent less staff; these same clients had been doubling their staff in the past five to seven years. To everyone’s surprise, in the first years, nothing happened! Doing less not only didn’t seem to impact the day-to-day business, it actually saved us a lot of money!
It was almost like delaying maintenance of your car. At first, there seems to be no difference in performance. After that, you may need to add some oil, but again, for a couple more months, the car simply gets you from point A to point B. But after awhile, little things start going wrong, and one day, you get stranded far from home.
This is exactly what happened in IT. The old “If it ain’t broke, don’t fix it!” attitude that some IT people have didn’t help here either. Upgrading software was one of the first things that suffered, and slowly but steadily, it’s becoming clear that this is a costly decision.
Making the Mainframe Run More Efficiently
Especially on the mainframe, both IBM and the Independent Software Vendors (ISVs) developing software to manage the mainframe have made great strides to support the latest hardware features. Most of the enhancements were implemented to make software run more efficiently and effectively. Support for specialty processors (such as the System z Integrated Information Processor [zIIP] and the System z Application Assist Processor [zAAP]), better memory management, support for more memory, and other technologies can all help run the mainframe more efficiently.
But to take advantage of these advances, it does require software updates. And it’s important to understand that to assure the stability and reliability of the mainframe, there’s no such thing as a simple upgrade. We need to set aside time and resources to make sure the new version of whatever product we install doesn’t introduce any surprises. But we’ve gone too far. Many IT organizations run their production environment on older versions of software, and often aren’t aware of both the advantages of the new software and the risk of continuing with older versions. Any upgrade requires a proper plan, including risks and benefits.
The Power of Upgrading to DB2 10
A good example of this is IBM’s DB2 for z/OS. Most of the mission-critical applications on mainframes run on this Database Management System (DBMS) and use a lot of expensive MIPS doing so. It also often serves as the back-end for many tablet and Internet applications. So, if you realize that the latest version offers an out-of-the-box performance improvement of 5 to 10 percent, you also realize that not upgrading simply means you’re wasting a lot of CPU time (which for most companies equals money).
Strangely enough, while some large mainframe shops have projects in place to tune applications and are trying to reduce the amount of MIPS, an upgrade to the latest version of DB2 isn’t considered. Of the companies that haven’t yet upgraded to DB2 10 (75 percent of all DB2 users), 25 percent are still on V8, which is no longer supported by IBM. Even those on DB2 9 will have to upgrade within the next 18 months before end-of-support for DB2 9. But why wait? Why keep on wasting so many expensive MIPS if you can free them up to do other things? Why not avoid “fines” when you cross the bar and pay extra for MIPS you only use because you temporarily need the extra power for end-of-year processing?
To make the upgrade process cheaper and more reliable, you will need the right tools. Using the right tools gives you the fastest path to value with the lowest possible risk. But there’s more. As noted previously, the workload on the mainframe has dramatically increased while the workforce has decreased. Is this one of the reasons why we’re wasting computer time by not upgrading? If so, we’re wasting valuable IT staff time as well. The 3270-character-based interfaces of the products your IT people work with to manage DB2 date back 20 to 25 years. Although these products have served us well, if we compare the complexity of DB2 today with that of 20 to 25 years ago, we must understand there’s a limit to what you can do with character-based interfaces.
Migrating a complex database today from one environment to another, with all the different components, is complex. So is analyzing performance over longer periods. The enormous amounts of data and events processed every day by our DBMS are so complex that we need to rethink the way we collect, view, and process information. Questions like: “I can see an abnormality in the past 24 hours, but is this a daily/weekly/monthly/quarterly/recurring pattern? And once I figure out that it is, how do I save that information so that when someone notices the same phenomena next time, no one spends time looking for the same answer? In other words, how do I prevent my IT staff from wasting time because the tools they have are simply no longer suited for some of the tasks I ask them to perform?”
When Gartner looked at the mainframe DBMS management market, they looked at more than just market share. Vision and power to execute were also important factors—and for the right reasons! When IBM introduced DB2 10 for z/OS, many ISVs were faced with the challenging task to support as well as exploit all the new capabilities this new version offers to help you stop wasting valuable time. Today, every company tries to do the right things to get out of the rough economic times in good shape. This includes running IT as lean as possible and making sure you run the software that helps you accomplish this. MIPS are expensive and so are your people. Making sure none of those two things are wasting any time on unnecessary things should be every manager’s top priority.