What's New In Data Building for Scale: AWS’s Marc Brooker on Distributed SQL
Mar 25, 2025
Marc Brooker, VP and Distinguished Engineer for Databases at AWS, dives into the world of Distributed SQL (DSQL) and its innovative serverless architecture. He explains how DSQL is reshaping cloud databases by balancing consistency, availability, and scalability, eliminating traditional hassles. Brooker shares insights on enhancing distributed systems with EC2 clock synchronization, achieving consistency in multi-region databases, and the exciting integration of AI with serverless technologies, all while setting new standards in high-performance cloud computing.
AI Snips
Chapters
Transcript
Episode notes
Customer Regret
- Many customers express "regret" after initially choosing easy-to-use databases.
- These databases often fail to scale as their applications grow, forcing costly re-architectures.
Consistency vs. Latency
- DSQL prioritizes strong consistency and durability over latency in its multi-region setup.
- It ensures data durability across two of three regions before acknowledging commits, increasing commit latency but preventing data loss.
Clock Synchronization
- DSQL leverages EC2's Precision Time Protocol (PTP) for clock synchronization, enabling strongly consistent reads.
- By referencing a transaction's start time (tstart), DSQL provides a consistent data snapshot without inter-region coordination.