A financial analytics company is migrating its historical transaction log archive () and an operational MySQL database from an on-premises data center to Google Cloud. The environment has an active Cloud Dedicated Interconnect connection to GCP. The migration plan requires completing the archive movement with minimal operational overhead, while minimizing downtime for the operational database during a planned 4-hour maintenance window. Which data transfer and migration strategy should the Cloud Architect recommend?
- Use Storage Transfer Service to transfer the log archive to Cloud Storage over the Dedicated Interconnect, and use Database Migration Service (DMS) with continuous replication to migrate the MySQL database to Cloud SQL.Answer
- BOrder a Google Cloud Transfer Appliance to physically transfer the log archive, and perform a manual database export using mysqldump to import into Cloud SQL during the cutover window.
- CUse Storage Transfer Service for the log archive, but migrate the operational MySQL database to Cloud Spanner using custom batch export scripts during the cutover window.
- DProvision an IPSec HA VPN tunnel over the public internet to migrate both the log archive via Storage Transfer Service and the MySQL database via Database Migration Service.
Answer
The optimal strategy is to use Storage Transfer Service to move the log archive to Cloud Storage over the Dedicated Interconnect, and use Database Migration Service with continuous replication to migrate the MySQL database to Cloud SQL.
Transferring 120 TB over a 1 Gbps Dedicated Interconnect takes roughly 11.6 days of continuous transfer, which fits standard online migration preparation timelines using Storage Transfer Service. For the 500 GB MySQL database, Database Migration Service (DMS) establishes Change Data Capture (CDC) replication from on-premises to Cloud SQL prior to cutover, allowing a zero-data-loss cutover well within the 4-hour maintenance window.
Step-by-Step Solution
Key Concept
Selecting Online Data Transfer vs. Physical Appliances and Continuous Database Replication Strategies