Datasets and results for Jelly-Patch benchmarks

Datasets and benchmark results comparing Jelly-Patch's performance to that of the formats built into Apache Jena. Benchmark datasets are available in two formats: .jellyp.gz (gzip-compressed Jelly-Patch) and .rdfp.gz (gzip-compressed RDF Patch text format). To run the benchmarks, you need the Jelly-Patch files. The .rdfp files are plain-text and can be easily inspected or reused in other benchmarks. Datasets: assist-iot-weather – the assist-iot-weather-graphs 1.0.3 dataset from RiverBench, transformed using a rolling diff. Each segment in the patch file corresponds to exactly one input frame (weather measurement), tracking the added and removed triples between each frame. Only the default graph was used to produce this dataset. bsbm-cdc – dataset generated using the RDF Delta change capture system, by running the Berlin SPARQL Benchmark 3.1 update workload on Apache Jena Fuseki. Each segment in the file corresponds to one database transaction. BSBM data generator parameters used: -fc -pc 100000 -tc 100000 -ud BSBM test driver parameters used: -runs 5000000 -ucf usecases/update/sparql.txt Benchmark code: https://github.com/Jelly-RDF/jvm-benchmarks Results: JSON files contain the serialization/deserialization throughput results. They follow the Java Microbenchmark Harness format. The size.tsv file contains the results of the serialized representation size experiment. License The datasets and results are licensed under CC BY 4.0. The original ASSIST-IoT dataset is also available under the same license. The BSBM data generator is available under Apache 2.0.