Incremental Materialization of RDF Graph Closures for Stream Reasoning

Incremental Materialization of RDF Graph Closures for Stream Reasoning
Alexandre Mello Ferreira
PhD Student

DEI - 3A Room
November 22nd, 2010


A crucial notion for Stream Reasoning is that of RDF streams, namely a stream of RDF triples annotated with a time-stamp, continuously produced by sources such as sensors placed in an environment. As it is clearly impossible to store a stream in its entirety, queries and reasoning tasks are formulated over RDF streams based on the notion of windows, i.e. the most recent triples in the streams, and extensively rely on the use of aggregation for extracting terse information out of verbose descriptions and for summarizing the past evolution of the stream.
The materialization of the transitive closure of an RDF graph has been extensively studied for the case of static knowledge, but the new scenario of dynamic knowledge, with triples flowing in and out of windows with time, calls for the need of new incremental inference techniques to add and delete consequences in the right order.
The research conducted in this field by the Database Group at Politecnico di Milano, to which the work reported here has contributed, moved from a survey of existing approaches to incremental inference for RDF to the proposal of a method that is specifically suited for the case of Stream Reasoning. More specifically, the contribution of the work described in this report is a demonstrator of the proposed technique applied to a power consumption monitoring scenario.

Alexandre Mello Ferreira

Research area:
Information Systems