Supporting Phase Management in Stream Applications

In modern streaming applications a tremendous amount of data is harvested, processed, and stored. Data stream or Complex Event Processing (CEP) engines are used to identify interesting, abnormal, or even dangerous patterns within these streams. In monitoring scenarios, however, often the temporal phases of the observed objects are interesting for the user together with the detection of abnormal patterns. In this paper, we argue that technical support for phase analysis allows for considerably reducing the complexity of continuous queries as provided, e.g., by the Continuous Query Language (CQL). Such phases provide an advanced layer of abstraction allowing for easily formulating phase-related queries in an intuitive way.