Migrating half a million vehicles and workloads to AWS cloud native platform
A customer in the automotive industry began their cloud journey in 2016 with a straightforward lift-and-shift migration from on-premises to AWS. This approach did not modernize applications and led to increasing operational inefficiencies. By 2024, an executive decision was made to modernize all services across 40 development teams to reduce rising costs and improve sustainability.
AWSMigratingServerlessThe Challenge
A major challenge was the presence of large components handling significant traffic volumes. Even when implemented in a microservices architecture, these services grew in size, accumulated technical debt, and slowed the delivery of new features. This created an imbalance between development and operations, further limiting agility.
Our Solution
With support from AWS, subject matter experts, and specialized AWS Partners, a new cloud-native target architecture was defined. Each development team was required to migrate workloads into a dedicated AWS account, enabling the use of the most appropriate services and tools for their needs. To accelerate migration, teams were provided with preconfigured CI/CD pipelines in AWS CodePipeline, including build stages in AWS CodeBuild and deployment stages in AWS CodeDeploy, supporting blue/green deployments. To simplify the transition, teams could deploy applications to Amazon ECS using either an EC2-based microservice pipeline or AWS Fargate to reduce infrastructure management overhead. Previously, teams relied on one Elastic Load Balancer (ELB) per component and large auto-scaling groups, driving up costs and limiting flexibility with savings plans. Under the new model, teams gained the option to adopt serverless architectures, enabling faster development cycles and lower operational burden for new solutions.
Results & Impact
Novalo Technologies supported one development team in migrating a service that handled 20% of the company’s traffic. The migration enabled the service to be deployed in Japan and South Korea, which was not previously possible. The cost savings were substantial: monthly AWS spend for the service dropped from $31,000 to $6,100 by leveraging ECS on Fargate, Lambda and a modernized target architecture. This success reinvigorated the team, fostering renewed interest in innovation and sparking a follow-up project to further modernize workloads and adopt serverless technologies where applicable.