GAIA: A System for Interactive Analysis on Distributed Graphs Using a High-Level Language
summary
GAIA compiles Gremlin into a distributed dataflow and ships a runtime that automatically parallelizes nested, dynamic traversals. It introduces the Gremlin-to-dataflow compilation and the Scope abstraction to group per-traverser state, checks dependencies at runtime to avoid per-traverser operators, and schedules batched tasks across cores. Deployed at Alibaba, it shows 64-node near-linear scaling, beats high-performance engines like Timely/JanusGraph with large speedups and lower memory, and is open source.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:30Graph Traversal
- 2:01Motivation
- 2:52Gremlin
- 5:21Scope
- 7:36Compilation
- 8:23Scope Extraction
- 9:49Efficient Implementation
- 12:11Evaluation Results
- 13:27Conclusion
related talks
37mGraph Computing at Scale
30mUsing Gremlin API and Apache Airflow to facilitate Complex System Integrations
34mApache AGE
01h 01mLogic and Algebras for Distributed Computing: Applications and Open Questions
30mScalable Graph-Parallel Inference Algorithms and Systems
31mApache Gravitino: A Multi-regional, Geo-distributed Meta Datalake
32mGlean: Data about Code
24mLarge Scale Graph Computation with Apache Giraph