Replatforming: Should you migrate from mainframe to cloud? Our expert opinion

The modernization of IT infrastructures It is key to improving the efficiency, security and scalability of business systems. 

And the options for achieving this are multiple and vary depending on the needs of each organization.

Modernizing may involve migrating a mainframe infrastructure to a cloud environment, as part of a project. replatforming. 

It may also include refactoring processes that involve a more in-depth restructuring of a system or application's code. 

Similarly, it may include the deployment of a microservices architecture and/or the process automationAnd it may even involve an upgrade of mainframe infrastructures, which, far from shrinking, continue to expand. 

In this article we analyze how to evaluate the possibility of developing a project replatforming, the implications and challenges it entails, and the most common mistakes that are made.

We also focus on the differences with refactoring, and the trends that will mark the future of replatforming.

To address these issues we interviewed Leo Zrycki, director of the IT Patagonia Mainframe Center of Excellence and quadruple IBM Champion zSystem, who provided us with his vision and perspectives to contribute to the development of this article.

What is replatforming?

In some cases, the replatforming It is a strategy that It involves moving applications from the mainframe to a new technological infrastructure, such as the cloud..

In order to be able to specify its definition, it is important to differentiate the types of projects that can be considered as replatforming. 

A specific case is move entire mainframe systems to Microfocus COBOLA change of environment, which means that programs will no longer be run on the mainframe, but in the cloud.

There are other projects that are closer to IBM's vision where The core of the applications remains on the mainframe, and the services in the cloud. 

IBM has a vision of a hybrid cloud, where a significant portion of the organization's processing goes through the mainframe (in COBOL), while some services in Java, Python and other programming languages, move to the cloud. 

The new versions of IBM's mainframe infrastructures support just those possibilities, and even the option of having Linux on the mainframe so that all modern languages such as Python, Java, and C can be run from it.

Furthermore, it is now possible to have DevOps on the mainframe. 

In fact, the IDz integrated development environment (IBM Developer for System Z), allows for graphical interfaces and working like any other modern language, where a programmer can see the database model and have help for system management.

In the first of the two cases, the mainframe is left behind. 

In the other, much of that infrastructure is leveraged and applications are left outside the mainframe.

Differences between replatforming and refactoring

He replatforming and the refactoring are application modernization strategies, but they differ in their scope and focus. 

He replatforming involves moving an application to a new infrastructure, starting from the completion of the changes needed to make it work in the new environment without significantly altering its architecture or code. 

Instead, he refactoring It includes a deep restructuring of the code to optimize its operation, improve its scalability or adapt it to new technologies, without modifying its external behavior. 

While the replatforming looking for speed and less impact on the business, the refactoring It requires more time and effort, but offers greater long-term benefits.

To illustrate this with specific examples, when moving from a mainframe COBOL to a cloud COBOL (Microfocus COBOL), the translation is almost straightforward. Almost, but not entirely, because everything related to the online environment can present differences. 

For example, there will no longer be platform-dependent screens through which bank staff view a transaction. 

In this case of replatforming, the screens that are developed with the software disappear CICS commands on the mainframe and they would have to be rewritten.

Refactoring, on the other hand, is about the constant improvement of programs. For example, when moving forward with program reengineering or when seeking optimal maintenance conditions so that anyone reading a program can understand it almost immediately.

Why are organizations considering replatforming?

The reasons that may lead a company to start a process of replatforming are multiple.

1) Economic reasons

Some companies are beginning to consider the possibility of migrating from a mainframe infrastructure to a cloud environment, understanding that this can provide a positive economic impact. This isn't necessarily the case. 

Before making a decision based on potential cost reduction, it's important to keep in mind that IBM's mainframes have incorporated new features. Now, Part of the processing can be done for a very small fee.

Indeed, there are coprocessors within the mainframe that allow you to run artificial intelligencePart of the database load, part of the mainframe Linux, runs on special zIIP processors, which are very economical. 

Therefore, there are also cost-effective options in the mainframe infrastructure space that should be evaluated to make the most favorable decision for each organization.

2) Lack of mainframe talent 

Another reason that may lead to the need to advance in an infrastructure migration is the shortage of mainframe and COBOL programmer talent. 

We are experiencing a situation where the demand for IT talent exceeds the available supply, which often requires hiring and training young people with little experience. 

Or, retrain people who are already working within the organization, so that they acquire new skills.

Without a doubt, the Lack of the number of professionals that the IT industry demands today It is a real bottleneck, and The industry needs answers to meet its own demand for employability.

These responses can be generated from the technology sector companies themselves, as is the case with IT Patagonia and its ally Codeki.

3) Market trend 

Companies are increasingly interested in migrating their applications, data, servers, and storage from traditional on-premise infrastructure to hybrid cloud services, or 100% cloud. 

By 2024, end-user spending on public cloud services worldwide was estimated to total $678.8 billion. This was reported by Gartner, noting that, compared to the $563.6 billion generated by the cloud market in 2023, the estimated figure for 2024 would represent an annual increase of 20.41% in 3Q.

As this trend evolved, many companies that had been working practically 100% on-premise with local physical infrastructures, They began to evaluate the migration of their resources to the cloud.

What are the most common mistakes organizations make in replatforming projects?  

A project of replatforming It is a process that extends over time, which can take months or even years, and which has a high risk threshold.

The investment in the software stock that organizations that work with the mainframe platform have is not easy to rewrite. 

In parallel, it is not easy to generate programs that run on other platforms and have adequate performance.

This calls into question the effectiveness and efficiency of this type of migration, which involves significant risks that must be properly assessed.

Therefore, the most common errors associated with this type of project are that a company:

• Press with a time acceleration.
• Do not measure the risks involved in the replatforming.

A perspective linked to the simplicity and speed of this type of process can generate false expectations and anxiety situations that threaten the success of the project.

In terms of risk assessment, carry out the correct testing of what needs to be done during the project and perform a constant monitoring of your performance, are necessary conditions to achieve results.

How to analyze the suitability of a replatforming project?

Based on the most common mistakes made by organizations that undertake projects replatforming, it is important to have analysis criteria that provide a sufficiently informed decision perspective.

From IBM They recommend that before undertaking any lift and shift migration, the factors that may affect the difficulty, cost, and final value of the project should be carefully evaluated. 

Among them, they mention the following:

– Application lifespan: It is important to define how much longer it will continue to be used or the current workload will continue to run. In this regard, they emphasize that in most cases It makes no sense to migrate an application that will be retired in the next 12 months..

– API Access RestrictionsIBM suggests ensuring that cloud migration doesn't create bottlenecks for current API tools.

– Migration priorityIf you plan to migrate multiple applications, it is recommended to create an execution manual to ensure that critical applications are migrated first (or in the order that makes the most sense for each business).

– Compliance: Evaluate the cloud infrastructure provider's migration plan and infrastructure to ensure that all compliance requirements will be met during and after the migration.

– Increased features and rangeFeature-rich cloud environments can be tempting to integrate capabilities immediately, but this can lead to delays and resource consumption. Having a clearly defined plan and adhering to it throughout the migration is critical.

Main challenges when migrating critical applications from the mainframe 

Always the most important challenge is performance. An absolutely critical point in any platform migration. 

The mainframe has very low latency, allowing organizations such as banks, retail, and telecommunications companies to process data 24/7 and execute a large volume of transactions simultaneously.

An example of speed is the speed of fraud detection with artificial intelligence, which a mainframe can resolve in tenths of a second with a chip.  

Another challenge is ensure that the process of replatforming do not compromise data protection and regulatory compliance.

In this regard, it's important to note that the type of hardware encryption used in IBM's mainframe infrastructure also takes quantum computing into account. The system is designed to be undecryptable, not even with quantum computing.

Conclusion: Trends that will shape the future of replatforming

After analyzing the implications of the projects of replatforming, the reasons that may guide its implementation and its differences with the refactoring, it is worth considering what trends will define it in the coming years.

In this sense, we can point out that the future of replatforming It will be defined by three key axes: modernization of the development cycle, performance optimization, and the use of new hardware architectures. 

• The modernization of the development cycle It involves adopting agile methodologies and DevOps tools that accelerate software delivery, reduce errors, and improve collaboration between teams. 
• The constant search for performance It will be critical to ensure that migrated applications not only run on the new infrastructure, but also leverage its capabilities to operate more efficiently and scalably. 
• In this context, the use of specific coprocessors and accelerators plays a central role, as it allows for increased computing capacity at a relatively low cost, optimizing intensive workloads and improving operational efficiency. 

Our Mainframe Center of Excellence directs its focus towards solving these types of projects and generating these types of solutions. Learn how we do it.