System Design · Uber · Hard
Design a storage system that serves large structured datasets to distributed compute frameworks, analogous to a data‑processing file system. The description must go beyond a box‑and‑line diagram: provide detailed workflows for metadata operations and data retrieval, caching mechanisms at multiple tiers, job dispatching, node crash handling, and workload distribution. Assumptions & Boundaries Characterize whether files are immutable, append‑only, or support random mutations;…
Checking your access…