Get tracked: a triple store for RFID traceability data

Companies are increasingly employing Radio Frequency Identification (RFID) technologies to track their goods in business processes. The rapidly produced, large amounts of events pose new challenges to modern databases: an efficient data staging process and efficient querying mechanisms for traceability data. In this paper, inspired by recent work on RDF triple stores, we present a scalable dedicated system for efficient storage and fast querying of RFID data. We address the challenges as follows: (1) we incorporate elaborated indexing techniques leveraging the specifics of the data in order to enable efficient event processing; (2) the query engine takes advantage of the RFID characteristics (e.g., the monotonic increase of timestamps) to speed up query processing. Our experimental studies show that the RFID Triple Store can achieve both a significantly higher insert throughput and a better query performance compared to a commercial row-oriented DBMS and the open-source column-store database system MonetDB.