A walkthrough of Twitter's system design covering core architectural concepts: user timeline, home timeline, trending hashtags, and search. Key topics include the fan-out approach for distributing tweets to followers' home timelines, handling celebrity accounts with millions of followers using a hybrid pull/push strategy, Redis-based caching for read-heavy workloads (600K reads/sec vs 600 writes/sec), stream processing pipelines (Kafka, Apache Storm) for real-time trending hashtag computation with geolocation mapping, inverted full-text indexing (EarlyBird) for search via scatter-gather across distributed nodes, and coordination via ZooKeeper for large Redis clusters. The database layer uses MySQL-based Gizzard and Cassandra for analytics.
•36m watch time
5 Impressions