How AWS S3 is built

AWS S3's massive scale and engineering are explored with Mai-Lan Tomsen Bukovec, VP of Data and Analytics at AWS, who has been running S3 for 13 years. S3, which holds over 500 trillion objects and serves hundreds of millions of transactions per second, manages tens of millions of hard drives across millions of servers. The discussion covers S3's evolution from eventual to strong consistency, detailing the replicated journal data structure and cache coherency protocol that enabled this shift without compromising availability. Also examined is S3's approach to durability, including the use of auditor systems to inspect every byte and the design principle of assuming constant failures. S3's roadmap is shaped by customer needs and a vision for making data usable, demonstrated by the introduction of S3 Tables and Vectors.