Skip to content

A benchmark for Read-Write Linked Data clients in a building automation scenario

Notifications You must be signed in to change notification settings

kaefer3000/rwld-brick-benchmark

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

96 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

rwld-brick-benchmark

A benchmark for Read-Write Linked Data clients in a building automation scenario. Simulates a building (IBM B3 of IBM Dublin) as described using the Brick ontology.

Prerequisites

  • Java >= 7. Tested with Java 8.
  • Maven >= Version 3.1
  • WGET
  • CURL
  • AWK
  • node.js
  • Free ports between 40100 and 40399
  • NxParser 3.0.0-SNAPSHOT installed:
$ git clone https://github.com/nxparser/nxparser
$ cd nxparser
$ mvn -Dmaven.test.skip=true install

We had to make a choice between Java 8 and Java 11 support for a dependency, currently we decided for Java 8 as it is still widely supported, but we will switch to Java 11 in the future.

Obtain

You have to tell git to also fetch the submodules git clone https://github.com/kaefer3000/rwld-brick-benchmark --recursive

Usage for rule-based automation benchmark

  • ./init.sh To initialise data
  • ./server.sh start all To start all servers
  • ./benchmark.sh To run the benchmarking script

Evaluation directions

  • Data access:
    • All data in one file on an HTTP server
    • Deployed as Linked Data: One-hop network around the IRIs from the materialised reasoning of Brick ontology + IBM B3 building description
  • Work loads:
    • Turn on the lights - baseline - baseline

    • Turn on the lights during working hours (eg. 7 am to 7 pm) - raising energy efficiency - adding conditions

    • Turn on the lights if weather report says it is dark - raising energy efficiency - adding conditions and external API

    • Turn on the lights if light sensors indicate - raising energy efficiency even more - many numerical computations

    • Turn on the lights if light sensors indicate (with individial thresholds) - adding individual preference for comfort - more numerical computations

    • Toggle the lights - adding conditions

Usage for workflow-based automation benchmark

  • ./init.sh -n <buildingcount> -h <hostname> To initialise data
  • ./server.sh -n <buildingcount> -h <hostname> -s <speedupfactor> start all To start all servers

Supply the number of buildings to be simulated, default: 1. If the client does not run on localhost, supply the hostname, default: localhost. The speedup factor sets the factor between wall clock time and simulated time, default: 613200.

Misc

  • If the a switch serves as manual override, it should have a neutral state
  • Find out where they say Brick uses RDFS-plus
  • The Turtle serialiser is the fastest of rdf-ext's serialisers

About

A benchmark for Read-Write Linked Data clients in a building automation scenario

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published