Mainframe Cost Reductions: Exploding a Few Myths
Reducing mainframe costs starts with understanding your systems, writes Peter James, capacity and performance management consultant, Vertali
IT and business managers are under pressure to reduce costs wherever they can, but without impacting the delivery of essential enterprise services. The good news is that by properly understanding the way you run workloads on your systems, and then optimizing them, it’s possible to wipe millions off your bottom line while remaining confident you are delivering the services you need. As IBM itself has advised, “Before you can get started reducing your mainframe costs, you should understand what drives those costs.”
The fact is, you have more control over your mainframe performance and costs than you may think. Analysis, infrastructure tuning and planning for capacity and performance improvements are the key—and you have multiple options.
When you combine mainframe profiling software with tried-and-tested modelling techniques, seasoned capacity management specialists can assess the efficiency with which workloads are processed by your systems. You can learn whether there is scope to, say, reduce the amount of CPU capacity required to process certain workloads, as well as highlight areas of concern or where more focused activity can deliver immediate benefits.
Taking Control of Mainframe Costs
There are many areas where costs can be optimized, laying to rest the fallacy that the mainframe is expensive to run, and is therefore poor value for money. Let’s look at a few simple examples.
One belief is that IBM software costs will simply go up and up. IBM actually offers a number of options to reduce your software bills by offloading some workloads to specially allocated processors.
Another fallacy is that it’s impossible to manage software costs driven by peak demand. Key IBM software products are charged based on a four-hour rolling average. Careful workload management means you can minimize the 4HRA to reduce software charges, with 10% being a typical target. It’s also possible, with careful planning, to control workloads to ensure they don’t exceed a predetermined usage threshold, thus controlling the costs.
It’s also incorrect that you can’t do anything about third-party software costs. Third-party (non-IBM) software is typically based on the size of the LPAR it is running on. Robust analysis and careful planning means it’s possible to set up LPARs of only the size that are needed for specific workloads, with the potential to reduce software costs significantly.
And we shouldn’t just accept that workloads will keep growing and growing, and that all we can do is add more capacity. It’s true that without regular analysis and tuning, workloads will tend to grow organically year on year. (It’s a bit like your home and contents insurance premium each year, simply accepting the renewal offer without checking first. Reductions are almost always possible). You need to take a good look at workloads regularly and investigate where savings in resource usage can be made. It’s an opportunity to reduce running costs while improving quality of service at the same time.
Informed Analysis Is the Key
These are just a few examples, and there are many more options available to optimize resource utilization with confidence while reducing running costs. Expert analysis is the key, with the results from your investigations driving next steps, such as:
- Reviewing existing capacity management processes to confirm they meet current and future business requirements.
- Confirming your existing approach to managing mainframe capacity allocation is optimal, including a review of LPAR configuration and workload management policies.
- Carrying out an analysis of workloads on other LPARs, as deemed appropriate.
- Performing a review of online transactional systems to confirm optimal configuration and looking for optimization opportunities.
- Identifying new capacity and optimization opportunities to minimize current requirements.
- Predicting additional capacity requirements ahead of time, so either deferring hardware changes or enabling acquisitions to be made in a more timely and therefore cost-effective manner.
The big prize is improving overall return on investment from your mainframe environment and, quite possibly, extending the lifespan of your current mainframe infrastructure to reduce costs even further.