Apache Beam is an open source, unified model for defining both batch and streaming data-parallel processing pipelines. Using one of the open source Beam SDKs, you build a program that defines the pipeline. The pipeline is then executed by one of Beam’s supported distributed processing back-ends, which include Apache Apex, Apache Flink, Apache Spark, and Google Cloud Dataflow.