BIFROST: A Future Graph Database Runtime
James Clarkson, Georgios Theodorakis, Jim Webber · Proceedings - International Conference on Data Engineering · 2024
BIFROST is a novel query engine for graph databases that supports high-fidelity data modeling on arbitrary and evolving graph topologies. It dynamically optimizes queries according to meta-level changes in the underlying graph (i.e. changes in topology) without the need for any explicit schema. This is possible by using state-of-the-art techniques from managed programming languages, such as self-optimizing ASTs and deoptimization, to combine query optimization and compilation. The approach provides high fidelity for even highly irregular labeled property graphs and gives good performance when compared to other systems that depend on fixed schemas for query planning and optimization.
Cited by 3 works. doi:10.1109/icde60146.2024.00448
Authors
References
- TensorFlow: A system for large-scale machine learning
- Pregel
- Group formation in large social networks
- MonetDB/X100: Hyper-Pipelining Query Execution
- Apache flink : Stream and batch processing in a single engine
- The dataflow model
- Efficiently compiling efficient query plans for modern hardware
- GraphX
- Cypher
- Proceedings of the 2019 International Conference on Management of Data