In this deep dive session from AWS re:Invent 2023, Grant McAlister, a senior principal engineer at Amazon Web Services, explores the innovative features and architecture of Amazon Aurora. He discusses how Aurora's unique design, which separates compute from storage, enables advanced capabilities like Global Database and low-latency read replicas. McAlister delves into recent innovations including Aurora I/O-Optimized storage, zero-ETL integration with Amazon Redshift, and improvements to Aurora Serverless v2. He also covers the addition of the pgvector extension for vector similarity searches to support generative AI applications. The presentation provides detailed insights into Aurora's internal workings, sharding capabilities, and how it addresses common database scaling and management challenges. McAlister demonstrates how Aurora's features can significantly improve performance, reduce latency, and simplify database operations for various workloads and use cases.