Map Reduce explained with example | System Design

3-Second Blueprint (English)

1. MapReduce splits large files into replicated DFS chunks and runs Map tasks locally on storage nodes to avoid network bottlenecks.
2. Mappers emit intermediate key-value pairs that the Shuffle phase partitions and groups by key across distributed Reduce workers.
3. A central coordinator tracks worker health and chunk locations, automatically re-executing failed Map or Reduce tasks for fault tolerance.