How much Semantic Data on Small Devices?
Download as PPTX, PDF0 likes630 views
The document analyzes the performance of semantic data tools on small devices, focusing on various benchmarks and the characteristics of small-scale ontologies. It compares results from different triple stores, revealing that Sesame generally outperforms others like Jena and Mulgara in loading time, disk space, memory consumption, and query time. The findings highlight that small-scale benchmarks can yield different outcomes than large-scale ones, emphasizing the need for further research on small-scale scenarios.
How much Semantic Data on Small Devices?
- 1. How much semantic data on small devices?Mathieu d’Aquin, AndriyNikolov and Enrico MottaKnowledge Media Institute, The Open Univeristy, UKm.daquin@open.ac.uk@mdaquin
- 2. Semantic Data on Small Devices?
- 3. Benchmarking Semantic Data ToolsLUBM(1,0)103,397 triplesLarge Scale Benchmarks
- 4. Extracting sets of small-scale ontologiesClusters of ontologies having similar characteristics, except for size
- 5. Extracting sets of small-scale OntologiesCharacteristics of ontologiesSize (tiples): varies from very small scale to medium scaleRatio class/prop: allowing 50% varianceRatio class/inst.: allowing 50% varianceDL expressivity: Complexity of the language99 automatically created clustersManual selection of 10
- 6. Results
- 7. QueriesUsing real life ontologies need domain independent QueriesA set of 8 generic queries of varying complexity, and which results might depend on inferenceSelect all instances of all classesSelect all comments Select all labels and commentsSelect all labelsSelect all classes (RDFS/OWL/DAML)Select all properties by their domainSelect all RDFS classesSelect all properties applied to instances of all classes
- 8. Running the benchmarks – Triple StoresJena with TDB persistent storageRAs above + RDFS reasoningSesame with persistent storageRAs above + RDFS reasoningMulgara with default configuration
- 9. Running the benchmarks – DeviceAsus EEE PC 700 (2G)
- 10. Running the benchmarks - MeasuresLoading time: for each ontologies in an empty, re-initialized store.Disk Space: of the persistent store right after loading.Memory consumption: of the triple store process right after loading the ontology.Query time: for each ontology, averaged over the 8 queries.
- 11. Results – Loading time
- 12. Results – Loading timeR=R
- 13. Results – Disk Space
- 14. Results – Disk Space=<<RR
- 15. Results – Memory consumption
- 16. Results – Memory consumptionsRR=
- 17. Result – Query time
- 18. Result – Query time=<RR
- 19. Conclusion – on testsSesame performs best in almost all aspects, even when including reasoningReasoning has big impact on Jena TDB at query timeMulgara is clearly not adequate in a small-scale scenario
- 20. Conclusion – on small-scale benchmarkingValidates our assumption that small-scale benchmarks give different results than large-scale benchmarksPoints out the need for more work to tackle the small-scale scenariosResults are not always clear cut in every aspects: benchmarks as support to decide which tool to use, depending on the application constraints