How Automated Refactoring can help you modernise mainframe applications onto AWS Cloud

01/12/2023 minute read Tim Jones

Legacy applications pose a significant challenge for organisations that rely on mainframes for critical business operations. Whilst COBOL, ADSO, CA Gen, Natural, and other pre-relational languages that support these applications are robust, their rigidity can impede technological innovation. These situations are further complicated by the fact that many developers skilled in these languages are nearing retirement age, leading to a shrinking talent pool and increasing maintenance costs.

There has been a growing interest in modernising to a cloud platform like Amason Web Services (AWS) due to various factors including the COVID pandemic, shifts in broader business strategies to address security concerns, and the advent of AI, among others.

The choice of approach for any organisation embarking on a modernisation project will depend on a range of factors. These include the desired technology stack, budget, timeline, available skills and resources, and the future trajectory of the organisation. For those organisations planning to completely transition away from mainframes, refactoring the pre-relational codebase to object-oriented languages such as Java or C# can drastically cut down operational costs. At the same time, it can mitigate the risks tied to an organisation's legacy systems.

What is refactoring?

Refactoring is a broad term that is sometimes used to describe an array of solutions, all focused on changing the original codebase to meet the organisation’s needs. This can be automated or manual, with transcoding and manual rewriting representing the two extremes.

Transcoding is an automatic process where a source language (like COBOL) is converted line-by-line into a target language (like Java), often referred to as "compile-time conversion". This method is inexpensive and highly automated, using licensed software to transform COBOL applications into Java for deployment on a standard Java application server. However, the resulting language – sometimes colloquially referred to as “JOBOL” – is often unmaintainable and heavily constrained, thus offering very few options for performance tuning, optimisation, and extension. On the other hand, manual rewrites are the most time and effort-intensive solution and result in new, stable, and error-free application code - but the scale, cost and complexity of this task are frequently underestimated. In addition, there is also the consideration that mainframe systems often lack adequate documentation, and their complexity and interconnectivity are generally underestimated. This lack of understanding often leads to significant scope creep, potentially resulting in disastrous outcomes. Striking a balance in the refactoring gradient is customised, rules-based automated refactoring, considered the ideal approach for modernising procedural languages such as COBOL to an object-oriented equivalent.

The benefits of AWS

Migrating and modernising your applications to AWS can significantly reduce the overheads associated with traditional mainframes. It eliminates the need for hardware maintenance and staffing costs, paving the way for significant cost savings. In fact, migrating mainframe applications to AWS can result in cost reductions of between 60% and 90%. This substantial saving opens up opportunities for customers to invest more in innovation, modernising applications that will lead to better business outcomes and greater agility.

The process of migration is comprehensive and manageable, with AWS and Advanced providing the necessary infrastructure, software, and tools to refactor and transform legacy applications with minimal disruption and downtime. We offer access to advanced analytic tools, which allows businesses to extract greater value from their data. This means that not only is data more accessible, but it also becomes a more powerful tool for driving business growth and development.

Migration can also be the foundation for transformation, improving operational models that result in greater business agility. Organisations can free up teams from undifferentiated tasks and refocus them to serve customer needs better. Additionally, moving away from proprietary mainframe technology allows access to a large pool of talented architects and specialists for design and operation. Reducing the risk of the mainframe retirement skill gap and attracting talent to modernise core business workloads.

Real-world example: The New York Times

The New York Times partnered with Advanced to modernise their traditional COBOL-based application into a contemporary Java-based application hosted on AWS. This project involved the automatic refactoring of the code and the data migration from old indexed-files to a relational database.

The mainframe application, originally known as CIS, was rebranded as Aristo after the migration. It provided essential business functions, including billing, invoicing, customer accounts, and delivery routing. Advanced was employed to refactor COBOL programmes into Java classes and migrate VSAM KSDS files to a relational database. This refactoring process ensured that the new Java application maintained the functional equivalency of the original COBOL application, and consisted of five stages:

  • Discovery: Advanced begins every modernisation initiative with a discovery, free of charge. Our modernisation experts work with customers to understand and document your business and technical goals and objectives. In The New York Times’ case there was mounting pressure to quickly lower operational costs, and a previous attempt to manually rewrite CIS had failed. Our Automated Refactoring approach promised functional equivalence, lower operational costs, and easier integration with modern technologies.
  • Assess and design: Before starting the transformation, The New York Times needed a deep understanding of their CIS environment to inform decision-making and identify potential challenges. An extensive Automated Assessment provided this insight, revealing unknown artifacts, relationships, unused assets, and components requiring special attention in the modernisation process. This step was crucial in identifying systems consuming over 3,500 data files produced by CIS daily and defining work packets for phased and agile modernisation.
  • Transform: The transformation phase involved automated refactoring, database modernisation, and creating the target operating environment, with work being carried out by both our team and The New York Times teams. Work packets were refactored, tested, and delivered into the DevOps pipeline in iterative cycles until the entire CIS system was transformed. This process included automated data migration from VSAM to a relational database and COBOL-to-Java refactoring. Simultaneously, the target environment and related operational infrastructure were established.
  • Test and deploy: This phase began with work packets delivered to The New York Times from iterative modernisation activities in the CIS transformation phase. If issues arose, the work packet was passed back to Advanced where we reviewed the code, adjusted our tooling, reprocessed the work packet, and delivered it back to The New York Times to resume testing and deployment. This process was repeated until the entire Aristo estate was deployed.
  • Support: We worked closely with The New York Times to ensure a smooth and error-free transition into production. Part of this transition included the cooperative construction and testing of a go-live production cutover plan to reduce the potential risks associated with application deployments.

Although the CIS-to-Aristo modernisation project began as a cost-cutting exercise, it led The New York Times to take methodical and incremental steps towards an ideal digital transformation through the adoption of innovative, cutting-edge technology. Alongside a 70% reduction in operational costs, the key result was a considerable improvement in customer service. This provided a competitive advantage in a unique industry that has experienced significant, dynamic shifts in its ecosystem in recent years. You can learn more about their success story in the full case study.

Looking to migrate to AWS? Advanced are an experienced AWS Technology Partner with expertise in mainframe modernisation. Get in touch to learn more about our mainframe to AWS assessment service.