Optimizing performance and saving MIPS in mainframe and hybrid environments

Mainframe environments continue to be fundamental pillars in sectors such as banking, insurance, retail and public bodies, thanks to their robustness, reliability and massive processing capacity. 

However, the growth in workloads and the increasing cost of MIPS (millions of instructions per second) make it necessary to rethink performance strategies. 

In this context, cloud and hybrid architectures emerge as key allies for optimizing operations and reducing costs.

In this article we analyze how to reduce MIPS consumption from the mainframe modernization.

We also discuss how to assess the feasibility of a change to optimize critical processes and evaluate MIPS savings scenarios based on infrastructure type.

In addition, we share the results obtained through process optimization.

What are MIPS?

MIPS (Millions of Instructions Per Second) is a metric traditionally used to estimate the processing speed of mainframes. 

Indica How many millions of instructions can a CPU execute in one second?, so it serves as a measure of raw hardware performance. 

In mainframe environments, MIPS are used as a technical reference, and are useful for sizing environments and comparing hardware generations.

As explained from IBM, HE They use MIPS to represent the capacity of the zSeries frame or LPAR. 

The measurement does not actually refer to millions of processor instructions per second. The value is actually derived from the figures published by IBM in the table. Large System Performance Reference (LSPR). The default is to use the single-image Processor Capability Index (PCI) value with the average Relative Nesting Intensity (RNI) for the hardware. 

What impact does the MIPS reduction have?

The reduction of MIPS in mainframe environments has a direct and significant impact, both in terms of operating costs and technological efficiency. 

Reducing MIPS consumption means the system can perform the same tasks—or even more—using fewer instructions per second.

This translates into lower CPU usage and therefore lower software licensing feesEspecially in charging models like IBM's, which are based on metrics like MSUs (Million Service Units), closely tied to MIPS consumption.

In other words, A reduction in MIPS use can represent thousands of dollars saved annually in organizations with high mainframe dependency.

In addition to the economic savings, Reducing MIPS helps optimize overall performance of the environment, freeing up capacity for new applications or processes without the need to expand the infrastructure. 

This is especially valuable in contexts where The mainframe coexists with cloud solutions, as it allows less critical loads to be redirected away from the central system and reserve intensive mainframe usage for core business processes. 

In sum, An effective MIPS reduction improves the sustainability of the mainframe environment and makes it more competitive against modern architectures.

Importance of reducing infrastructure costs in a context of high transactionality 

Three key challenges facing organizations are reducing operational costs, preventing system outages, and improving and optimizing customer service. 

In a context experiencing exponential growth in transaction rates as people seek to operate with home banking 24/7; and in which we have become accustomed to conducting all types of transactions and purchases online, These challenges are becoming increasingly central.

Regarding cost reduction, it's important to understand how much a transaction consumes in terms of machine resources, as this impacts monthly maintenance costs.

IBM charges for the service of keeping the mainframe operational, calculating costs and billing using two formulas: average or peak consumption. 

Therefore, it is very important have transactions and processes running efficiently. 

How to evaluate the feasibility of a change to optimize critical processes? 

We share some steps to identify the pain points that conspire against the goal of improving the performance of processes and transactions:

1. Become aware of the problem

Knowing what problems affect an organization's performance is the first step toward optimizing processes and moving toward cost reduction.

Although identifying the problem is not enough, it is also essential understand what the real impact is on the customer and the systems.

For example, delayed batch processes reduce the availability of a home banking platform and prevent a person from performing certain transactions during certain nighttime hours, which may be most convenient for them.

This creates a sense of discontent among customers, which, while not easily measured, can lead them to begin wanting to change banks. 

To become truly aware of and understand the dimension of the problem and the need for solve it with an optimization, is the first step. 

2. Investigate what is happening 

The second step should be to focus on understanding the current situation, both in online and batch processes. Both are equally important. 

We know that an online process or transaction can take place in just a few seconds, but if executed multiple times, it can lead to a loss of performance in terms of the usability of an application used by the customer. 

The same thing happens with a batch process that takes a long time or is cancelled, in the case that it is a process that is already optimized or under certain programming rules in green code.

The realization of a assessment adequate ensures that the batch process is as healthy as possible and that, with proper planning, it can run on time and in the correct manner. 

This will allow progress to be made in solving the problem and effectively meeting service level agreements (SLAs).

3. Analyze critical pain points

Once the problem has been detected and made aware of, solutions must be taken Identify and analyze the pain points that this problem generates. 

At this point, it is necessary to determine how to approach the solution, starting from an inventory of everything that must be observed, in order to:

• Prioritize critical points, taking into account that the priority list is not static and must be adapted to the evolution of consumption.
• Work on the analysis of the causes that generate this delay or poor performance of systems, transactions or processes.

There are many technological tools that can be used as Omegamon, Six Performance Analyzer u OPTI, an exclusive product that we developed in IT Patagonia.

4. Consolidate work teams

The fourth phase involves the formation of highly trained work teams and with a long history.

Having professional teams allows organizations to ensure the continuity and efficiency of their systems and avoid negative impacts on their operations and security.

5. Evaluate the feasibility of change

Often, transformations are proposed that are very expensive to implement and whose cost-benefit ratio is not optimal. 

For example, it is thought or estimated that a certain change will significantly improve performance, but when its feasibility is evaluated, it is concluded that the benefit obtained is not that great. 

When an organization considers that it must make a change in a transaction, because it understands that it will cause a decrease in process consumption, it is necessary to justify it. 

How is this justified? In a laboratory environment, various executions of transactions or processes are performed, with the production version and the modified version.

From these tests, it is possible to determine whether a small change can generate a positive result, or whether a large modification does not produce the expected results.

MIPS savings scenarios by infrastructure type

As we explained before, the efficient management of MIPS is a constant priority for organizations that operate mainframe systems.

Depending on the type of infrastructure adopted - whether mainframe-based or hybrid - there are different scenarios and strategies to save on MIPS without compromising the performance or security of the systems. 

We present the main approaches and technical considerations for each case.

1. 100% mainframe infrastructure

In addition to what we noted previously, to reduce MIPS consumption without affecting service, the most effective strategies include:

• Refactoring COBOL or PL/I code to eliminate redundancies and improve efficiency.
• Using zIIP (z Integrated Information Processor) to offload eligible processes and free up overall capacity.
• Batch process tuning, adjusting execution windows and load prioritization to avoid unnecessary peaks.

2. Hybrid architectures (mainframe + cloud)

The key is to achieve a seamless integration between both environmentsThis requires a strong API governance strategy, service orchestration, and unified monitoring. Latency and data security must also be carefully managed.

Best practices in this scenario include:

• Smart Offloading from non-critical processes to the cloud, such as reports, secondary batch processes or validations.
• Exposing mainframe services through APIs to easily integrate with cloud applications.
• Use of distributed platforms to run tasks outside the z/OS environment without affecting the core business.

3. 100% cloud infrastructure

The transition requires a profound reengineering process: not just migrating code, but rethinking processes, data, and architecture. 

In addition, legacy applications must be refactored or replaced, which involves considerable technical and organizational effort..

To get the most out of this model, it is essential to:

• Redesign application architecture with a cloud-native approach (containers, microservices, DevOps).
• Adopt managed services that reduce operational burden and improve time-to-market.
• Implement continuous testing to ensure quality in dynamic environments and frequent changes.

What are the results achieved through process optimization?

Evidence shows that optimized processes can achieve a reduction of at least 20 to 30% in usage times.

On the other hand, choosing the correct compilation options from COBOL 6 onwards can reduce CPU usage to between 7% and 10%.

This data is of great importance, considering that the goal is to reduce transaction times for people, to avoid delays and cancellations, among other situations.

Having a lean process means they consume fewer MSUs and reduce machine load. This ensures peace of mind for both the client, who will be able to operate with ease, and for the operational or development department, who will be confident that their processes are running correctly. 

Conclusion

Optimizing performance and saving MIPS is not just a technical issue. It is also strategic.

Taking a holistic approach that combines mainframe best practices with the power of the cloud can improve efficiency, reduce costs, and lay the foundation for a sustainable technological transformation.

You may be interested in: Why core modernization is key to the evolution of the financial system.