Adaptively Brokering Continuous Queries for Exploiting Streams

This paper introduces an approach for adaptively brokering queries for exploiting streams produced and consumed by devices with limited computing and storage capacities. Such devices can be mobile and static and they can build networks for exploiting streams by evaluating queries in a collaborative way. Query brokering strategies are based on the notion of relevance which depends on application requirements and execution contexts used to associate priority to queries and streams. Since query brokering can be done in dynamic environments, relevance can change and thus brokering strategies can be dynamically adapted. The paper illustrates the feasibility of the approach by showing how it can be customized for three different application scenarios: a system for observing industrial environments, a system to share events (e.g., accidents,available parking spaces, emergency vehicles) in a vehicular network, and a P2P application to compare prices of products in different stores.