Past Research Projects

Partial order driven stream processing

Links: Flumina on Github Star, Diffstream on GitHub Star, Dependency-guided synchronization paper (PPoPP 2022), Synchronization Schemas (Invited at PODS 2021), Diffstream paper (OOPSLA 2020)

Collaborators: Rajeev Alur, Filip Niksic, and Caleb Stanford

Existing abstractions for stream processing consider streams to be either totally ordered sequences, completely unordered relations, or some fixed point in between, e.g., sequences of relations. However, these representations create several issues related to performance, too much order does not expose available parallelism, and correctness, lack of order disallows expressive temporal queries. To address these challenges, we propose a flexible partial order abstraction that can capture fine-grained ordering requirements, allowing for correct and maximally parallel stream processing.

Our work is open source and available on Github (Diffstream, Flumina) and if you want to learn more about it start from our Synchronization Schemas paper that describes types for partially ordered streams.