Migration from monolith to microservices with legacy compatibility

Today's ISR (Intelligence, Surveillance and Reconnaissance) defense coalitions require storage and disseminationmechanisms that are able to cope with emerging changes to requirements and new features. Previous Systemof systems (SOS) architectures used to be built with years of planning, development, testing and deployment,usually in the form of distributed monoliths. Due to new requirements in ISR, shorter response cycles arerequired. To reach this goal, new approaches are of interest in the architectural style and workload sharingwithin the development team, resulting in the ability to better maintain and change existing software solutions.Ideally, such a shift results in improved scalability, replaceability, modularity and resilience.In this context we examined our existing software that provides and also internally uses legacy middlewaresuch as ""Common Object Request Broker Architecture"" (CORBA) (among others). The overall codebase waswritten in such a manner that it was easy to produce, i.e., technically motivated. The development team israther small, so efficiency and the possibility to share (developer) knowledge is important.Our goal was to evaluate the state of the art, thus being able to reasonably apply modern software developmentapproaches that support mandatory legacy support. We attempted a restructuring of the codebase applying theprinciples of ""Domain-Driven Design"" with its ""bounded contexts"", resulting in domain-oriented source codethat is easy to verify and maintain.Keeping in mind our small development team, we aimed for shared responsibility, giving us the necessaryresilience for unplanned staff absence.In this publication, we present a possible migration path with its operational constraints (e.g., legacy interfaces) towards a more suitable software solution and the lessons learned during the process. In addition, weoutline how this was achieved with a small headcount.