Increasing Scalability and Reducing Costs Using Amazon Aurora Serverless with BMW

BMW is a global manufacturer of luxury cars and motorcycles and a provider of premium financial and mobility services. It includes storied brands such as Rolls-Royce, MINI, and BMW Motorrad. The company needed to scale its customer messaging system, BMW Messages, to handle more than 10 million messages per hour during peak periods. BMW Messages connects the entire fleet of BMW vehicles with customer care services to provide quick solutions for customer requests, over-the-air software update announcements, and key servicing information. BMW sought to improve its live messaging availability and database storage flexibility.

For the backend of BMW Messages, the company needed to process and retain vast amounts of data, consisting of transaction logs between company systems and customer vehicles. The stored data was used for customer support, delivering notifications for missed calls, service requirements, and reminders. The data is also used for legal purposes; for example, in the event of traffic incidents, the company must provide detailed data to government entities. In some countries, BMW needs to retain up to 10 years of historical data, so it was important to find a scalable solution to help manage its mission-critical data.

To improve scalability, BMW migrated its on-premises databases to Amazon Aurora, a relational database management system that is fully compatible with MySQL and PostgreSQL and combines the speed and availability of high-end commercial databases with the simplicity and cost-effectiveness of open-source databases. To seamlessly complete the migration, BMW used AWS Database Migration Service (AWS DMS), a managed migration and replication service that helps migrate database and analytics workloads to AWS quickly, securely, and with minimal downtime and no data loss.

BMW chose Amazon Aurora Serverless v2—which scales instantly to hundreds of thousands of transactions in a fraction of a second while adjusting capacity in fine-grained increments to provide the right amount of database resources based on application needs—to mitigate wasted resources. To account for usage spikes, BMW previously had to run workloads at maximum settings throughout the day, which incurred significant costs. Using Aurora Serverless v2, BMW has automated management of its database capacity and can scale to accommodate peaks on demand.

“We noticed a disproportionate increase in demand for scalability versus what we had seen in the past,” says Marc Fiedler, product owner for BMW Messages at BMW. “So, we decided to take a serverless approach on AWS to scale flexibly and, therefore, reduce costs.”

Previously, BMW had dedicated operations teams at each BMW hub around the globe. Each team was responsible for monitoring system performance daily by tracking workloads and database scaling issues. After migrating to Amazon Aurora PostgreSQL-Compatible Edition, the department could reassign 12 team members to focus on new product design.

“When we use Aurora Serverless v2, the amount of effort required to manage the database is drastically reduced,” says Fiedler. “Our developers can now allocate their time to furthering innovation and adding more features.”

Throughout the migration, BMW turned to AWS as it optimized all systems. For instance, the team needed to install security patches whenever it made a change in the database kernel. BMW deployed across multiple Availability Zones to minimize disruptions between patches, but failing over from its primary to its secondary database could take 1–5 minutes—a downtime BMW wanted to remove fully. So, the team reached out to its AWS solutions architect and implemented zero-downtime patching. Since then, using Aurora Serverless v2, BMW has not experienced any outages or downtime.

“Having downtime was a big pain point for us from a technical and legal perspective because it had the potential to affect important customer interactions,” says Fiedler. “After we submitted the feature request, we turned to the AWS solutions architect team and resolved this issue very quickly.”

After migrating to AWS, the company also unified its global architecture—consisting of 1,300 microservices that process more than 12 billion requests daily—and improved international connectivity and interoperability. BMW found that it could create consistent architecture for BMW Messages and other products using Aurora Serverless v2 in all the company’s hubs. In 6 months, BMW rearchitected its related systems around the world to keep its international processes running smoothly and scaling efficiently.

“We’ve moved away from the limitations of physical servers and are scaling in the cloud worry-free with Aurora Serverless,” says Fiedler. “Whenever we want to spawn new products in a serverless or cloud environment, we know we can work with a capable AWS team to get technical support and design ideas.”

Following the success of the BMW Messages migration to Amazon Aurora PostgreSQL and use of Aurora Serverless v2, BMW plans to move the messaging service toward a fully serverless architecture. By expanding its solutions to be serverless, the company can continue to optimize costs and work to improve the reliability of all its systems.

“Our mission is to go serverless and scale to enhance the user experience,” says Fiedler. “Using AWS, we have the solutions to make it possible.”