What is Apache Spark and how does it differ from Hadoop MapReduce?

Apache Spark is an in-memory distributed processing engine, while Hadoop MapReduce is disk-based batch processing system.

Spark improves performance by keeping intermediate computation results in memory (RAM), reducing disk I/O, which is a major bottleneck in Hadoop MapReduce. Hadoop writes intermediate results to disk after each map and reduce phase, making it slower. Spark supports DAG execution, lazy evaluation, and multiple workloads (batch, streaming, ML, SQL), whereas MapReduce is strictly batch-oriented.