Skip to content

Quickly Configure Iguana

Here we will setup a quick configuration which will benchmark one triple store (e.g. apache jena fuseki) using one simulated user. We assume that your triple store (or whatever HTTP GET endpoint you want to use) is running and loaded with data. For now we assume that the endpoint is at http://localhost:3030/ds/sparql and uses GET with the parameter query

Further on the benchmark should take 10 minutes (or 60.000 ms) and uses plain text queries located in queries.txt.

If you do not have created some queries yet, use these for example

SELECT * {?s ?p ?o}
SELECT * {?s ?p ?o} LIMIT 10
SELECT * {?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> ?o}

and save them to queries.txt.

Your results will be written as an N-Triple file to first-benchmark-results.nt

The following configuration works with these demands.

# you can ignore this for now
datasets:
  - name: "Dataset"

#Your connection
connections:
  - name: "Fuseki"
    # Change this to your actual endpoint you want to use
    endpoint: "http://localhost:3030/ds/sparql"

# The benchmark task
tasks:
  - className: "Stresstest"
    configuration:
      # 10 minutes (time Limit is in ms)
      timeLimit: 600000
      # we are using plain text queries
      queryHandler:
        className: "InstancesQueryHandler"

      # create one SPARQL Worker (it's basically a HTTP get worker using the 'query' parameter
      # it uses the queries.txt file as benchmark queries
      workers:
        - threads: 1
          className: "SPARQLWorker"
          queriesFile: "queries.txt"

# tell Iguana where to save your results to          
storages:
  - className: "NTFileStorage"
    configuration:
      fileName: "first-benchmark-results.nt"

For more information on the confguration have a look at Configuration