When is the Peak Performance Reached? An Analysis of RDF Triple Stores

With significant growth in RDF datasets, application developers demand online availability of these datasets to meet the end users’ expectations. Various interfaces are available for querying data using SPARQL query language. Studies show that end-points may provide high runtime performance at cost low availability. For example, it has been observed only 32.2% public endpoints have a monthly uptime 99–100%. One possible reason this is workload experienced by endpoints. As complete execution performed server side (i.e., endpoint), processing result degradation or even service shutdown. We extensive experiments capabilities well-known triple stores their In particular, we stressed with multiple parallel requests from different agents. Our revealed maximum after which point they lead shutdowns. hope analysis will help store design workload-aware engines improve throughput.