AWS DMS migration to Oracle RDS

The client had a large on-premise Oracle database on aging hardware, which was hosted in a data center that was targeted for decommission. The client’s IT strategy was to rehost to AWS. The database size was large (over 20TB), and the workload was highly transactional, both of which complicated migration planning. A more complex migration strategy was required to accommodate the nature of the platform

CleanSlate evaluated multiple Proof of Concept (POC) approaches to determine the best solution. This included creating an on-premise export of the database and utilizing AWS Snowball to transfer the data between the data center and AWS S3. Ultimately, CleanSlate completed a design which best met the clients needs which included:

• AWS Oracle RDS as a target database to simplify the overall management and data replication of the database
• Utilization of AWS Data Migration Service (DMS) to synchronize the on-premise database with an AWS Oracle RDS instance running in the cloud. This allowed the cloud version of the database to act as a near real-time replica of their on-premise database
• Implementation of custom scripts to produce real-time record counts and additional data validation
• Leveraged MuleSoft as the common ESB to continuously move/migrate Data from the older system(s) to a new system(s)
• Enabled CI/CD processes and continuous development/continuous integration

CleanSlate successfully completed the database migration with the following results:

• Cutover activities were executed in just minutes as DMS kept the AWS Oracle RDS instance synchronized with the on premise database.
• AWS Data Migration Services allowed the database migration progress to be completed “behind the scenes” and reduced the need to involve key client resources from other projects.

Project was completed within the expected schedule without unplanned customer downtime or outages. Client experienced zero data loss or data integrity issues.