The dataflow model

TLDR

Unbounded, unordered, global‑scale datasets are increasingly common, yet consumers demand event‑time ordering, feature‑based windowing, and rapid responses, making it impossible to fully optimize correctness, latency, and cost and forcing practitioners to adopt disparate implementations. We argue that data processing must shift from attempting to materialize complete data pools to embracing the reality of continuously arriving, retractable data, enabling principled trade‑offs among correctness, latency, and cost. The Dataflow Model provides such principled abstractions, defining semantics that support continuous data streams and offering core design principles that guide implementation choices. Our validation demonstrates that the Dataflow Model successfully supports these semantics and trade‑offs in real‑world deployments, confirming its practicality.

Abstract

Unbounded, unordered, global-scale datasets are increasingly common in day-to-day business (e.g. Web logs, mobile usage statistics, and sensor networks). At the same time, consumers of these datasets have evolved sophisticated requirements, such as event-time ordering and windowing by features of the data themselves, in addition to an insatiable hunger for faster answers. Meanwhile, practicality dictates that one can never fully optimize along all dimensions of correctness, latency, and cost for these types of input. As a result, data processing practitioners are left with the quandary of how to reconcile the tensions between these seemingly competing propositions, often resulting in disparate implementations and systems. We propose that a fundamental shift of approach is necessary to deal with these evolved requirements in modern data processing. We as a field must stop trying to groom unbounded datasets into finite pools of information that eventually become complete, and instead live and breathe under the assumption that we will never know if or when we have seen all of our data, only that new data will arrive, old data may be retracted, and the only way to make this problem tractable is via principled abstractions that allow the practitioner the choice of appropriate tradeoffs along the axes of interest: correctness, latency, and cost. In this paper, we present one such approach, the Dataflow Model, along with a detailed examination of the semantics it enables, an overview of the core principles that guided its design, and a validation of the model itself via the real-world experiences that led to its development.