Extending Sliding-Window Semantics over Data Streams

Abstract

Data stream processing is now commonplace in applications such as network monitoring, sensor networks, telecommunications data management, Web personalization, manufacturing and others. The continuous sliding-window query model is used widely in data stream management systems. However, the existing sliding window query models fail to answer some of the queries that qualify a certain condition. In this paper, we extended the existing sliding window to general scenarios by adding a new class of sliding window operator, termed condition-based sliding window. The condition can be defined over any attribute of data stream tuple in an out of order manner. We discuss the semantics of the operator and show that above method performs well for queries that qualify a certain condition.