Which architecture excludes the batch layer, processing data solely as streams?

The Kappa architecture is specifically designed to process data entirely as streams, effectively excluding the batch layer that is present in other architectures, such as the Lambda architecture. In Kappa architecture, all data is treated as a continuous flow, allowing for real-time processing without the need for batch jobs. This is impactful because it simplifies the data processing pipeline by focusing on a single stream-processing layer.

In contrast, the Lambda architecture incorporates both batch and stream processing layers, which can lead to increased complexity. Delta architecture also utilizes both batch and streaming data but focuses on data versioning and updates, while stream architecture, although potentially related to streaming data, does not specifically denote a recognized model distinct from Lambda or Kappa. Thus, Kappa architecture is the clear choice for environments that prioritize real-time data ingestion and processing without the complexities of a batch system.