Fidelity-based continuous query introspection and adaptation

Abstract

In this paper, we address the fidelity of information collected from distributed, interactive, pervasive computing environments. Queries in these dynamic networks come in two forms: snapshot queries and continuous queries. The former executes through the network and returns values that satisfy the query criteria at a particular instant in time. The latter expects a result that is updated over time to reflect changes in the system. The fidelity, or quality, of the results of these queries can be significantly impacted by the unpredictability of the dynamic network; as the query is distributed and executed, changes that occur due to network and environmental dynamics can cause the query to miss potential results or to report inconsistent information. Our previous work has shown how we can provide semantically meaningful statements about the fidelity, or quality, of the result of a snapshot query in a dynamic networked environment. In this paper, we demonstrate how this fidelity can also be used to adapt a continuous query's processing to make its execution best match the conditions in the underlying dynamic environment.