RWE Supply & Trading Speeds Database Migration to Amazon Aurora by 6 Months Using Amazon DMA

Headquartered in Germany, RWEST is the interface between RWE and the energy markets around the world. RWEST needed to modernize its trading systems to create greater operational resiliency and increase scalability to better adjust to an ever-changing global marketplace. Achieving the performance RWEST needed using its previous database infrastructure—large, monolithic, complex databases hosted in on-premises data centers using commercially licensed software—would be operationally challenging and costly. It would need to modernize its database solution to a nimbler microservices architecture.

RWEST decided to implement an open-source solution built on PostgreSQL and turned to Amazon Aurora, a relational database designed for unparalleled high performance and availability at global scale with full MySQL and PostgreSQL compatibility. In addition to saving costs by breaking free from commercial licenses, RWEST chose Aurora because it offered high performance and scalability of up to 15 read replicas. Aurora also offered operational resilience with six copies of data across three availability zones, continuous backups with point-in-time recovery, and failovers in less than 10 seconds.

Together, AWS and RWEST assessed 11 applications totaling 19 TB of data and decided to focus initially on migrating five workloads to Amazon Aurora based on customer priority. Due to the corporate-wide prioritizing of work on renewable energy and RWEST engineers’ inexperience with PostgreSQL, company stakeholders evaluated whether they should manage the migration themselves or bring in a third party. “The opportunity to migrate to Amazon Aurora PostgreSQL involved rearchitecting the applications and databases, and we couldn’t cover all of that in house,” says Dr. Tobias Bluhm, chief architect for RWE Supply & Trading.

RWEST and Amazon DMA used AWS Schema Conversion Tool (AWS SCT)—a migration tool that automatically assesses and converts source database schema and code objects to a format compatible with the target database—and AWS Database Migration Service (AWS DMS)—a managed migration and replication service to migrate database and analytics workloads to AWS quickly, securely, and with minimal downtime and zero data loss. By working with Amazon DMA, RWEST saved time and manual resources and simplified its database migrations using automation and the breadth of database expertise that Amazon DMA provided.

The Amazon DMA team assessed RWEST’s databases and applications to develop a road map for the migration. In September 2021, the project began with the migration of smaller, simpler databases as a proof of concept. RWEST and the Amazon DMA team worked in close collaboration. They shared information about migration paths, future state architectures, and detailed migration plans, and created a runbook with information about the deployment of each in-scope application and target database. “By getting runbooks, migration scripts, and hand-over sessions from Amazon DMA, our teams spend 50 percent less effort on the migration with less risk and faster cutovers,” says Bluhm. The Amazon DMA team educated and shared expertise with RWEST engineers during the migration. As the migration progressed, RWEST engineers gained skills, knowledge, and capabilities. “Our teams saw that we could do it,” says Bluhm. “They became more confident.”

Gradually, RWEST and the Amazon DMA team increased the complexity of the workloads that they migrated. The collaboration was so successful that RWEST accelerated the process and moved up the scheduled migration of a complex database by 6 months. RWEST saved additional time by using the automated tooling of AWS DMS to alleviate the burden of basic manual tasks such as data transfer and schema replication. “Our internal teams can concentrate on reacting fast to market changes and building new features, which requires a lot of market knowledge,” Bluhm says.

In addition, RWEST saw cost benefits because it can adjust database specifications to the criticality of the workload. “Using AWS DMS is not just one size fits all,” says Bluhm. “I really like that AWS offers a service that you can consume and that makes it possible to find the resilience you need while still balancing the costs.”