Semantic Publishing Benchmark v2.0

Introduction

The Semantic Publishing Benchmark is an LDBC benchmark which measures the performance of RDF databases. Creation of that benchmark has been inspired by the Media/Publishing industry where requirements like: constant update of content or various requests for data extraction and aggregation are present.

Features of the benchmark:

• Provides a Data Generator using real reference datasets and producing synthetic data of various scales.
• Workload consists of editorial operations (adding new, updating or deleting existing data) and aggregation operations (retrieve content, aggregate results, etc.). Aggregation operations consist of a wide range of queries, e.g. search, aggregation, FTS, Faceted Search, Geo-spatial, Drill-down which define various types of choke points (i.e. technical challenges) that an RDF database must successfully overcome.
• Provides validation of results
• Tests the conformance of the RDF Database to various rules inside the OWL2-RL rule-set.

Build

Apache Ant build tool is required.

# builds the benchmark driver with basic query mix, standard SPARQL 1.1 compliance
$ ant build-basic-querymix

# builds the benchmark driver with advanced query mix, standard SPARQL 1.1 compliance
$ ant build-advanced-querymix

# builds the benchmark driver with basic query mix and queries optimized for GraphDB
$ ant build-basic-querymix-graphdb

# builds the benchmark driver with advanced query mix and queries optimized for GraphDB
$ ant build-advanced-querymix-graphdb

# builds the benchmark driver with basic query mix and queries optimized for Virtuoso
$ ant build-basic-querymix-virtuoso

# builds the benchmark driver with advanced query mix and queries optimized for Virtuoso
$ ant build-advanced-querymix-virtuoso

Result of build process is saved to a distribution folder: 'dist/' :

• semantic_publishing_benchmark.jar - the benchmark test driver
• data/ - folder containing all necessary data to run the benchmark
• test.properties - a configuration file with parameters for configuring the benchmark driver
• definitions.properties - a configuration file with pre-allocated values used by the benchmark. Not to be modified by the regular benchmark user.
• readme.txt

Install

All necessary files required to run the benchmark are saved to folder: 'dist/'. The benchmark can be started from there or can be moved to a new location.

Configure

Various properties are used to configure the behaviour of the SPB Test Driver or Data Generator. All properties are saved in files: test.properties and definitions.properties. Properties saved to file: definitions.proeprties are not to be modified by a regular user of the benchmark, their default values have been set.

• RDF Repository configuration

  • Rule-set RDFS(subPropertyOf, subClassOf), OWL(TransitiveProperty, SymmetricProperty, SameAs) - see file RdfsRules-optimized-spb.pie in the distribution for details
  • Enable context indexing
  • Enable text indexing (optional)
  • Enable geo-spatial indexing (optional)

• Benchmark Actions are the essential tasks that the benchmark driver can perform e.g. Generate synthetic data, Validate operations or measure performance.
  Descriptions of the essential becnhmark actions can be found here: https://github.com/ldbc/ldbc_spb_bm/wiki/Benchmark-Actions. Each action consists of a set of operational phases (https://github.com/ldbc/ldbc_spb_bm/wiki/Operational-Phases) that are executed sequentially.

• Configuration options: https://github.com/ldbc/ldbc_spb_bm/wiki/Configuration-Options

• Definition properties: https://github.com/ldbc/ldbc_spb_bm/wiki/Definitions-Properties

Run

java -jar semantic_publishing_benchmark-*.jar test.properties

Note: appropriate value for java maximum heap size may be required, e.g. -Xmx8G

Benchmark Results

Logging details can be controlled by a configuration file: log4j.xml saved in the distributed benchmark driver (semantic_publishing_benchmark.jar). After modifying log4j.xml, benchmark driver must be updated with contents of the new xml file. Results of the benchmark are saved to three types of log files :

• brief - brief log of executed queries, saved in semantic_publishing_benchmark_queries_brief.log
• detailed - detailed log of executed queries with results, saved in semantic_publishing_benchmark_queries_detailed.log
• summary - editorial and aggregate operations rate, saved in semantic_publishing_benchmark_results.log