SOFOS: Demonstrating the Challenges of Materialized View Selection on\n Knowledge Graphs

Analytical queries over RDF data are becoming prominent as a result of the\nproliferation of knowledge graphs. Yet, RDF databases are not optimized to\nperform such queries efficiently, leading to long processing times. A well\nknown technique to improve the performance of analytical queries is to exploit\nmaterialized views. Although popular in relational databases, view\nmaterialization for RDF and SPARQL has not yet transitioned into practice, due\nto the non-trivial application to the RDF graph model. Motivated by a lack of\nunderstanding of the impact of view materialization alternatives for RDF data,\nwe demonstrate SOFOS, a system that implements and compares several cost models\nfor view materialization. SOFOS is, to the best of our knowledge, the first\nattempt to adapt cost models, initially studied in relational data, to the\ngeneric RDF setting, and to propose new ones, analyzing their pitfalls and\nmerits. SOFOS takes an RDF dataset and an analytical query for some facet in\nthe data, and compares and evaluates alternative cost models, displaying\nstatistics and insights about time, memory consumption, and query\ncharacteristics.\n