A media company stores petabytes of objects in Amazon S3 and has enabled S3 server access logging into a separate logging bucket for security audits. The security team runs ad hoc Amazon Athena queries against these logs a few times per month to investigate suspicious access patterns. Over the past year, the raw text-based access logs have grown to hundreds of terabytes, and Athena query costs and scan times have become unacceptably high because each query scans enormous volumes of uncompressed data. The team wants to reduce Athena query cost and improve query performance while retaining all historical log data for compliance. Which approach BEST meets these requirements with the least ongoing operational effort?