Arne Gerhartd, Deloitte Consulting; Scott Radeztsky, Deloitte Consulting
Keywords: Application Modernization, Legacy Transformation, Code Conversion, Automation, Refactoring, Natural, ADABAS, FORTRAN, Assembler
Abstract:
Goal of research: The goal of Deloittes research is to develop a fully automated process that can convert legacy code used on many SSA systems (e.g. FORTRAN, Natural, Assembler), databases, and screens to object-oriented languages (e.g., Java and C#/.NET) and Oracle as well as other relational databases and distributed data platforms. This solution would result in modernized systems that function exactly the same as the legacy system, but at a fraction of the cost and timeline of a from-scratch effort. This approach would minimize impacts to the no-fail missions, operators, interface partners and IT operations teams. Also, it would produce modern Java or .NET (C#) code that is familiar to any current technical team because it is similar in flow/structure as well as naming conventions. The learning curve for developers becoming productive in the new environment is reduced.
Methodologies used: Our transformation approach uses a unique, automated, risk-reduced, and proven methodology for converting mainframe applications to Java or.NET. Our automated conversion approach works based on two main principles:
One-to-One. The structure (source code and application screen modules) of the legacy application platform is converted to a new platform on a one-to-one basis with the fewest possible changes to achieve an accurate conversion of the legacy system functionality (e.g., critical algorithms). This approach reduces risk of missing existing functionality and accelerates testing and validation as the resulting system is identical in function to the familiar legacy system. This process is unique and uses patented technology.
Automated Conversion. Source code and data are transformed automatically in a repeatable way using a predefined schema. The process is consistent and produces the same result each time it is performed because any errors that are identified are addressed centrally in the transformation process. This feature allows for the parallel operation of both the legacy and refactored systems during implementation without the need for a code freeze, as modifications to the legacy system can be automatically migrated to the new environment through the same process.
Expected outcomes: Our approach to legacy transformation focuses on methodical approach to gauge and attack the legacy technology baseline while avoiding the risks. Additional benefits of the approach would include:
Capability to convert databases, code, and screens to modern languages/technologies. Once converted, resident capabilities of modern technologies can be used allowing for debugging, system expansion, advanced analytics, and scalability to accept more inputs/data sources.
Parallel system functionality (legacy and modern) and modular deployment. Extensive testing practices to validate data inputs/outputs. Methodology inherently prevents data loss.
No limit on performance, in fact, the modernized system can take advantage of modern infrastructure and tools to meet and exceed the speed and capacity needed to handle current data loads that is critical to integrated SSA capabilities.
Once converted, resident capabilities of modern technologies can be used allowing for debugging, system expansion, advanced analytics, and scalability to accept more inputs/data sources.
Applicability to field of SSA/SDA: Across the SSA/SDA portfolio, mission critical systems can be found running on a legacy technology stack. These systems may have existing performance constraints, may not be able to easily leverage advancements in technology (i.e. cloud, AI/ML, DevSecOps), and are reliant on dwindling institutional knowledge and select vendors, and requiring higher costs to maintain. Given many systems in the SSA/SDA portfolio are well beyond end of life, still operating because of the criticality of the mission and function, or simply because prior costly modernization efforts have failed, our solution and approach minimize risk and provide assurance that mission processing can occur without degradation.
Date of Conference: September 19-22, 2023
Track: Space Domain Awareness