Evaluation of performance of the RML Mapping Service
Approach
We randomly generate fifty queries that call the \emph{RML mapping service} to transform raw data of one country from the World Bank data source into RDF and fifty queries that invoke this service to transform raw data of all countries into RDF.
In addition, we use eight queries to transform raw data of the UK data source into RDF.
We calculate the time between invocation of the RML mapping service and the generation of the corresponding output RDF data set.
We run each query three times and calculate the average time needed to finish the query.
Note that after each query, we wait 60s before sending the next query to avoid the RML mapping service sending queries in an avalanche to servers of the World Bank and the ONS.
Details
Queries that transform data of a country into RDF |
List 1 |
Queries that transform data of all countries into RDF |
List 2 |
Average time for data transformation of a country |
Result 1 |
Average time for data transformation of all countries |
Result 2 |