
Continental improved autonomous vehicle simulation performance by 20% and reduced costs by 30% using Mountpoint for Amazon S3 to directly access data without data movement orchestration, enabling use of general-purpose EC2 instances instead of high-I/O specialized hardware.


<div><p>This article describes how Continental uses Mountpoint for Amazon S3 to improve autonomous driving simulation performance by 20% and reduce costs by 30%.</p><ul><li>Mountpoint for Amazon S3 eliminates need for data movement orchestration between S3 and compute</li><li>Enables switching from high-I/O EC2 instances to general-purpose instances, reducing costs</li><li>Processes 3,000-5,000 recordings (500 TB total) in parallel using AWS Batch array jobs</li><li>Two integration variants: dynamic S3 locations per container or static location on EC2 host</li><li>Removes 15-minute average data transfer overhead per simulation run</li><li>ADDF framework extends EKS module with pre-installed Mountpoint for S3 support</li></ul><p>Mountpoint for Amazon S3 simplifies Continental's ADAS/AV simulation architecture by providing direct S3 access without complex orchestration, improving performance and cost efficiency.</p></div>


Related articles