Skip to content
FindOpenSource

Beam

Unified programming model for defining batch and stream data processing pipelines that can run portably across multiple execution engines.

Overview

Apache Beam lets a team write a data pipeline once using its unified batch/stream API, then run that same pipeline on different underlying execution engines - including Flink and Spark (also in this catalog) - without rewriting it for each. This avoids locking a pipeline's logic to one specific processing engine.

Beam is a fit for teams that want pipeline portability across execution engines, or who are building on Google Cloud Dataflow (which uses Beam as its programming model), a different layer than Flink or Spark themselves, which are the engines Beam pipelines actually run on.

Categories
Analytics
Keywords
data-pipelinebatch-processingstream-processingportable-pipelines
Languages
Java, Python, Go
License
Apache-2.0

Spotted an error? Suggest an edit on GitHub.