The authors of the influential Dataflow Model paper reflect on their VLDB Test of Time award-winning work, evaluating what aged well and what was missed over eleven years. While core principles like event time primacy and strong consistency proved sound, they acknowledge that windowing/triggering dominated unnecessarily and that the stream-centric view missed a deeper insight: streams and tables are two representations of the same object. The paper traces how database mechanisms like SQL, incremental view maintenance, and materialized views ultimately delivered on the original analytical goals more effectively.
Background
The Dataflow Model paper is a foundational work in streaming analytics that introduced concepts like windowing, triggers, watermarks, and retractions for processing unbounded data. This retrospective evaluates its lasting impact and evolution in the context of the VLDB Test of Time award.
- Source
- Lobsters
- Published
- Sep 8, 2026 at 01:11 AM
- Score
- 7.0 / 10