External server replication in Cloud SQL is a configuration that replicates data from a self-managed source database into a Cloud SQL replica. The source sits outside Cloud SQL, which is what separates this from the internal read replicas you create entirely within the managed service. The Professional Cloud Database Engineer exam tends to test whether you can recognize when this feature applies and what the two main scenarios for it are, so it is worth being precise about what it does and does not cover.
The feature supports replicating data from a self-managed source into a Cloud SQL replica. It is useful when the source is not already a managed Google Cloud service, which means you are dealing with a database you run yourself. The source can live in Google Cloud, in another cloud provider, or on premises. Whether the primary database runs on a virtual machine in another provider or on a physical server in your own data center, you can connect it to Cloud SQL and have changes flow into the replica.
External server replication is available for MySQL and PostgreSQL. It replicates an external source into the corresponding Cloud SQL version of that engine, so a MySQL source replicates into Cloud SQL for MySQL and a PostgreSQL source replicates into Cloud SQL for PostgreSQL.
The same replication mechanism supports two different goals, and telling them apart is the part the exam cares about.
The first is database migration. Here you connect your source database to a Cloud SQL replica in order to move the workload into Cloud SQL with minimal downtime. While replication runs, your application keeps pointing at the original source, and the data on the Google Cloud side stays continuously updated. When you are ready, you perform a cutover to Cloud SQL with very little interruption to the service. The replica exists as a stepping stone, and the destination is Cloud SQL.
The second is an ongoing hybrid setup. In this case the source database remains the primary point of entry for the application, and replication runs continuously to extend the data into Google Cloud. This fits a situation where you need to keep the primary database in its existing environment but still want read or processing capacity on a replica in Google Cloud. There is no cutover here, because the external source stays in charge.
The distinction that matters for the Professional Cloud Database Engineer exam is the direction of intent. With migration, the external server is temporary and you are heading toward Cloud SQL as the new home. With a hybrid setup, the external server stays primary and the Cloud SQL replica is a permanent extension of it. Both use the same external server replication feature, so a scenario question often comes down to whether the goal is to leave the source behind or to keep running on it.
Our Professional Cloud Database Engineer course covers external server replication alongside Cloud SQL read replicas and high availability, with practice questions that drill these distinctions.
