Tag: sanjay-ghemawat

Google's 2003 File System paper described a distributed storage system designed around commodity-machine failures, enormous files, streaming reads, and append-heavy workloads, making failure handling and scale first-class properties of the file system rather than exceptional conditions.

Jeffrey Dean and Sanjay Ghemawat's MapReduce turned a narrow pair of user functions into a contract with a distributed runtime that handled partitioning, scheduling, data movement, and machine failure, allowing ordinary programmers to use thousands of cluster machines without writing a distributed scheduler.