Scalable execution of continuous aggregation queries over web data

Abstract

Data delivered over the Internet is increasingly being used to provide dynamic and personalized user experiences. Queries over fast-changing data from distributed data sources are executed to create content to be delivered to users. Because these queries require data from multiple sources, they're executed at intermediate proxies or data aggregators. The authors discuss various techniques for executing aggregation queries over distributed data to minimize the number of message exchanges between data sources, aggregators, and users. They carefully examine the problem in terms of different types of queries, aggregation functions, query imprecisions, and whether the aggregators get data from sources using pull- or push-based mechanisms.