MiniSpark

MiniSpark executes MapReduce-style transformations on a single node using worker threads, a global task queue, locks, and condition variables. The project focuses on scheduling partition-level work as dependencies become available and materializing a DAG of transformations in the correct order.