WatDiv Extended Version

Introduction

(work-in-progress)

WatDiv benchmark comes with a set of 20 query templates. agnos.ai has been working on extending the complex query (C) to cover SPARQL features that are commonly used in more realistic business-like use case scenarios. This is a work-in-progress and more queries are planned to be added in the future.

The benchmark can be run against many dataset sizes. This particular benchmark was run for 100M-triple (15.6GB on Disk).

Description

The execution of the WatDiv Benchmark includes the following steps:

1. Set up instances of the triplestore to be tested

2. Load one single n-triple file with 100 million triples to an empty triplestore database

3. Run a sequence of customised WatDiv C1 type queries

4. Run a very large update immediately followed by selects of the modified data

5. Run a select count version of each query variation to force the graph patterns to be resolved against the full query dataset and not only a limited number of triples

6. Update triplestore statistics and re-run test queries