Many of the applications that need to be migrated are sitting on top of a monolithic architecture without well-defined boundaries and a large number of data interdependencies between systems. The client decided to divide the migration into phases in order to move pieces of the monolith at different points in time. However, this implies that some applications will access the data from the new platform while other applications will access the same data from the legacy platform.

Data synchronization was implemented to capture new updates from the source database platform early in the project to meet specific business objectives. However, historical data loads were not part of the initial rollout, and these are required to fully migrate any application. Real-time synchronization limits the ability to perform initial data loads because the database needs to remain online.