A Partial Evaluation Approach for Querying Data Streams and Distributed Fragmented Relations

Due to the significant communication overhead incurred in the presence of a rapidly updated stream, centralised approaches are becoming less realistic for large scale distributed processing, especially when complicated continuous queries such as window aggregation and joins between streams and relations are processed. This paper proposes techniques to evaluate stream queries in more generic distributed systems, based on a novel application of partial computation. The efficiency and effectiveness of the proposed techniques and algorithms have been verified by an experimental study.