Design a Distributed File System
Hard · System Design · about 45 minutes
Asked at Google, Databricks, Snowflake, Amazon, Microsoft, Meta, CrowdStrike.
Design a distributed file system like GFS or HDFS that stores ~10 PB of mostly large files across thousands of commodity servers. It should give batch jobs high-throughput sequential reads and concurrent appends, and survive the disk and machine failures that happen every day at this scale.