Crystal
Compressed execution approach for distributed subgraph matching on very large graphs.
Crystal is a method for distributed subgraph matching on very large graphs that computes compressed results directly instead of materializing large intermediate outputs.
What it is
Crystal targets the output crisis in subgraph matching: for many graph pattern queries, materializing all matches or intermediate matches can dominate the overall cost. Crystal studies compressed representation and computation for such workloads.
Key ideas
- Compress subgraph matching results to reduce output materialization cost.
- Compute compressed results directly instead of first materializing all matches.
- Reduce intermediate result overhead in distributed graph pattern queries.