zkSPARQL-bench

Using the bechmark

The benchmark is generated and released to

https://registry.npmjs.org/zksparql-bench/-/zksparql-bench-X.Y.Z.tgz

The benchmark can be directly downloaded and used:

• the dataset can be found in the credentials contained in dist/BENCHMARK_NAME/data-signed/
• the documentLoaderContent.json gives all URL resolutions that need to be overridden when performing the benchmark

Contributing

A benchmark for zero knowledge SPARQL queries over Verifiable Credential. To generate the benchmark run

npm i
npm run build

Requirements:

• Docker
• Nodejs (>= 22.x)

SP2B queries

Fetch and generate the SP2B query files into sp2b/queries:

npm run sp2b:queries

They will be copied into dist/sp2b/queries during npm run build.

TODO:

• Add a DCAT description of the benchmark
• Add query answer checkers
  • For LUBM this can be based on https://swat.cse.lehigh.edu/projects/lubm/
• Check the generated queries
• Dont publish the dist to github, it publishes each file individually which is time consuming

Known Limitations:

• Credential subjects are not always well modelled in order to suit the benchmark. For instance reviews should not be the subject of a credential.