Rivian Executes Vision of Agile Engineering on AWS

To meet accelerated engineering schedules and reduce the need for physical prototypes, electric vehicle manufacturer Rivian relies on advanced modeling and simulation techniques. Using high compute capacity, simulations enable engineers to test new concepts and bring their designs to market quickly.

In 2020, Rivian found that its on-premises research and development information technology infrastructure could not keep up with its performance needs. Resource bottlenecks affected product lifecycle management, computer-aided design, and computer-aided engineering, so Rivian began using Amazon Web Services (AWS) to architect an agile engineering environment.

Rivian depends on computer-aided engineering tools to extend vehicles’ range and maintain high safety standards. But in early 2020, one of the company’s on-premises high-performance computing clusters failed, reducing its compute capacity by half. Rivian looked to the cloud to overcome this challenge.

“Our engineers expected the fix to take 6 months,” says Madhavi Isanaka, Rivian’s chief information officer. Instead, Rivian built a new compute cluster on AWS. “In 3 weeks, we had a working proof of concept on AWS,” says Isanaka. After that success, the company migrated its production environments. In the cloud, Rivian’s engineers can access and automate resources on demand.

On AWS, the speed of Rivian’s software tools has improved by up to 66 percent, and Rivian can load a full vehicle bill of materials in 22 minutes. The company uses Amazon Elastic Compute Cloud (Amazon EC2) C5 Instances, which deliver cost-effective high performance at a low price per compute ratio. By using Amazon EC2 C5n Instances and Elastic Fabric Adapter (EFA), a network interface for Amazon EC2 instances, Rivian’s engineers can scale out to a larger number of cores.

AWS Select Consulting Partner X-ISS provides system and application technical support to Rivian’s computer-aided engineering team for Scale-Out Computing on AWS, which helps customers deploy and operate multiuser environments. “In early product development stages, we don’t have many physical vehicles, so we use AWS to bring the design space to life,” says Isanaka. Using Amazon FSx for Lustre, a fully managed storage service, Rivian can access shared storage quickly. And after consulting AWS Professional Services, a global team of experts, Rivian improved data availability using Amazon EC2 Auto Scaling, which helps users maintain application availability.

On AWS, interaction with product lifecycle management has increased 66 percent. Rivian also improved failover using Amazon Relational Database Service (Amazon RDS)—which makes it simple to set up, operate, and scale a relational database in the cloud. Backup synchronization, which before took up to 1 day, now takes less than 1 hour. “As Rivian grows at a rapid pace, we need a highly scalable system,” says Surendra Balu, Rivian’s 3DExperience technical lead. “Changes that took 5 days now occur within minutes.” And using AWS CloudFormation, which enables users to speed up cloud provisioning, Rivian can deploy automatically through continuous integration / continuous delivery.

Rivian plans to continue migrating workloads to AWS, enabling more seamless postprocessing and visualization. “People who were skeptical about high-performance computing in the cloud are more open minded after seeing our results on AWS,” says Isanaka. “This is accelerating adoption across the board.”