chaozhang-cs / bic

This repository provides the BIC index, an index for processing graph connectivity queries within sliding windows on streaming graphs.

Bidirectional Incremental Computation (BIC)

This repository provides the BIC index, an index for processing graph connectivity queries within sliding windows on streaming graphs.

Organization

baselines includes the implementation of the state-of-the-art indexes, compared in the experimental study.

bic includes the implementation of the BIC index.

benchmark shows the code for experimental evaluation.

Getting Started

1. Clone the project.
2. Execute mvn clean package.

Reproducibility

1. Download the datasets_and_workload.tar.gz, including the datasets and workloads used in the experimental study, and place the tar file under the benchmark folder.

2. Execute tar -czvf rlc-benchmarks.tar.gz.

3. benchmark/datasets and benchmark/workloads contains the datasets and workloads.

4. Execute nohup java -Xms768g -Xmx950g -cp target/swc-1.0-SNAPSHOT.jar ca.uw.dsg.swc.benchmark.BenchmarkRunner & to reproduce the results. This can take weeks to complete all experiments. Make sure JDK 11 or higher is installed on the tested server. In case the server does not have enough memory for running experiments with extremely large graphs, like Friendster, comment on the settings with such graphs in BenchmarkRunner.

5. The benchmark results are available under the directory benchmark/results.

Contact

Chao Zhang, [email protected]