An initial investigation into the performance model (profile) of Google Cloud DataFlow which provides a simple, powerful programming model for building both batch and streaming parallel data processing pipelines.