The Dataflow Model: A Practical Approach to Balancing Correctness, Latency, and Cost in Unified Stream Processing, Tyler Akidau, Robert Bradshaw, Craig Chambers, Slava Chernyak, Rafael J. Fernandez, Joseph C. Hoe, Hannes Kiefer, Stephen Kirby, Alan Markmaker, Francis O'Donovan, Sam Robb, Martin Rosenbach, Eric Schmidt, Vadim Shamis, Keith Turner, Robert Vawter, 2015Proceedings of the VLDB Endowment, Vol. 8 (VLDB Endowment)DOI: 10.14778/2824032.2824076 - A foundational paper introducing the Dataflow Model, which establishes key concepts like event time, watermarks, and various windowing strategies essential for accurate and robust stateful stream processing.
Apache Flink Documentation: Stateful Stream Processing and Time & Windows, The Apache Flink Community, 2024 (Apache Software Foundation) - Official documentation sections detailing Apache Flink's robust state management, fault tolerance mechanisms (like checkpointing), and advanced event-time processing capabilities including watermarks and various window types.
Apache Kafka Streams Developer Guide, Apache Software Foundation, 2024 (Apache Software Foundation) - The official developer guide for Kafka Streams, covering its core concepts, architecture, and how to implement stateful stream processing applications, particularly relevant for Kafka-centric architectures.
Feast Documentation: Ingesting Data and Online Serving, The Feast Community, 2025 (Feast Project) - Official documentation explaining how Feast, a popular open-source feature store, handles data ingestion, including from streaming sources, and how features are served for real-time model inference.