Distributed-systems

Notes on Discord's migration from Cassandra to ScyllaDB and their data modeling strategy for handling trillions of messages with strict latency requirements.

Notes on how Meta solved the cache invalidation problem at scale using a system called Polaris that guarantees cache consistency without sacrificing performance.

Notes on how Cloudflare scaled Kafka to handle over 1 trillion messages daily across their analytics and logging infrastructure.

Notes on Google's software load balancer that handles traffic for Search, YouTube, and other services using consistent hashing with minimal disruption.

Notes on Google's seminal 2004 paper that introduced a programming model for processing massive datasets across thousands of commodity machines.