Switching replication engines with zero downtime is a topic with immense complexity. In this paper, we discuss how a data replication engine can be changed or upgraded without taking either the application or the database offline. There are numerous methods to thwart common problems faced during a migration: Version Independence (which avoids interoperating different software versions); the challenges when utilizing two nodes; utilizing three nodes (for partial hardware migrations); utilizing four nodes (for a full hardware refresh); and bi-directional environment (that keeps all nodes synchronized as the migration takes place). Additionally, the so-called jagged edge problem must be resolved, which occurs when performing any replication engine migration. HPE Shadowbase software offers several methods to solve these problems, including HPE Shadowbase Zero Downtime Migration, which can upgrade applications, database schema formats, file and table locations or indices, operating systems, and can refresh hardware. Properly undertaken, a data replication engine migration imposes no downtime on applications or users, and the databases all remain accessible, consistent, complete, and up-to-date during the process.