vladimircaribb - stock.adobe.com
When it comes to making mainframe operations efficient, it is important to consider that each workload running on a corporate mainframe is tied to a specific business process. As such, IT must demonstrate that the workload is generating an ROI and that the cost of running the workload is not higher than or disproportionate to the revenue the associated business process generates.
Given that mainframe workloads and their associated costs have a direct impact on an organization's bottom line, it is essential to keep mainframe costs in check. At the same time, however, it is also important to ensure the reliability of your mainframe workloads. After all, as expensive as it might be to run a workload, an outage is likely to be far more expensive.
To put it another way, those who are responsible for mainframe operations management must maintain a delicate balance of making sure that workloads are reliable and that they receive the resources they need to function properly, while also avoiding any unnecessary costs. Fortunately, there are some strategies to optimize your mainframe operations and make them efficient.
Make processes more efficient
Unfortunately, there is no magic formula for improving the efficiency of mainframe operations. Efficiency ultimately comes down to the hardware and the software working together in an optimal manner. That being the case, optimization efforts might initially focus on making sure that hardware resources are not being wasted nor overstressed.
IBM Mainframes, for example, generate a CPU activity report you can use to get a better understanding of how your available CPU resources are being consumed. (You can find IBM's guide to the CPU activity report on its documentation webpage.)
When reviewing this report, you might begin by looking at the
LPAR/MVS BUSY TIME PERC value. If the value is high, it could mean there is contention for CPU resources. You can confirm this by checking the
DISTRIBUTION OF IN-READY WORK UNIT QUEUE. If the value is higher than 60%, CPU contention is likely occurring. Conversely, if the
LPAR/MVS BUSY TIME PERC value is low, it may mean that bottlenecks exist within the system and that the CPU resources are not being used as efficiently as possible.
Of course, not every performance problem is hardware-related. A performance problem can just as easily stem from poorly written code. While there are various utilities you can use to look for inefficiencies within your code, the process of sifting through mountains of code can be both tedious and costly. You might narrow things down by looking at historical performance data to see which applications' performance have diminished over time. Once you have identified such applications, it may be worth seeing if you can correlate drops in performance to any actions taken -- such as moving the application, upgrading the application to a new version, an increase in the application's workload or, perhaps, a change made to a seemingly unimportant subcomponent.
Reduce operating costs
Oftentimes, improving operational efficiency will translate directly into reduced costs. However, there may be other ways to reduce your costs.
One possible option is to perform a cost assessment on each application to see if it might cost less to run a given application in the cloud.
If you would rather keep your applications running on your own mainframe, then your best option might be to take a hard look at your licenses and any contracts -- licensing contracts, support contracts and so on -- that you have in place. Perhaps your organization has been paying to renew contracts it no longer needs. Likewise, you might discover that making a simple change to the way some of your resources are licensed could save you a considerable amount of money.
Another way that you might reduce your operating costs is to look for redundant applications. If, for example, you have two different applications that do essentially the same thing, you might be able to reduce your costs by consolidating around a single application. Even if this approach does not end up reducing your licensing costs, it will likely reduce your support costs.
Decrease the risk of downtime
Another area of focus for those interested in operations management is to look for ways to reduce the risk of downtime. There are countless ways in which an organization might improve application availability, from installing backup generators to building a failover data center in another region.
ROI is the most important consideration when it comes to taking steps to reduce the risk of downtime. For each application, determine the hourly cost of an outage. Once you know how much an outage will truly cost the organization, you will have a better feel of how much you should realistically be spending on risk mitigation measures. After all, redundancy tends to be quite expensive, and you will most likely be asked to justify any investments made. If you're able to contrast the cost of such an investment against the cost and likelihood of an outage, you can build a more compelling case for investing in downtime mitigation solutions.