Modern distributed systems need identifiers that are unique, scalable, efficient, and often sortable. In this talk, Mons Anderson, Solution Architect at Exness, will trace the evolution of identifier generation — from database autoincrement and UUIDs to Snowflake, KSUID, NanoID, ULID, and other modern schemes.
The talk will compare these approaches by uniqueness guarantees, sortability, storage efficiency, readability, operational complexity, and suitability for distributed systems. Special focus will be on UUID v7, which builds on lessons from previous generations and is becoming a new industry standard.