Apache Hadoop and Apache Spark are two leading frameworks for distributed big data processing that have significantly impacted geospatial analytics. Both systems use clusters of commodity hardware in a shared-nothing architecture to scale out horizontally, allowing massive spatial datasets to be processed in parallel. Hadoop popularized the MapReduce programming model and excels at batch processing of very large files. Spark is a newer engine that builds on some of Hadoop’s concepts but introduces in-memory data processing and a more flexible execution model, often yielding faster performance for many tasks. This entry focuses on the differences between Hadoop’s disk-based MapReduce approach and Spark’s in-memory approach, especially in the context of spatial (vector and raster) data processing. We also highlight several systems that extend Hadoop or Spark specifically for spatial data, and discuss emerging trends toward integrating big data frameworks with higher-level query processing.