How AppsFlyer modernized their interactive workload by moving to Amazon Athena and saved 80% of costs

This article discusses how AppsFlyer modernized their Audiences Segmentation product by migrating from Apache HBase to Amazon Athena, resulting in significant performance improvements and cost savings.

Specifically, the article covers:

• AppsFlyer's Audiences Segmentation legacy architecture using HBase and the challenges faced with scalability and performance
• The decision to move to Amazon Athena for better cost-effectiveness and performance
• Schema design optimizations in the new Athena architecture, including partition projection, data sorting, and schema refactoring
• Performance optimizations like parallel query runs, using Parquet file format, and query result reuse
• The testing, monitoring, and rollout process for the new Athena-based solution
• Improvements achieved, including a 10% increase in query performance, 80% cost reduction, and enhanced system stability
• Conclusion highlighting the benefits of using Athena and the various optimization techniques employed