A technique where changes are written to a sequential log (InnoDB’s redo log) before being applied to data pages. WAL guarantees durability and enables crash recovery — MySQL can replay the redo log after an unexpected shutdown to restore data consistency.