Incorporating Support for Event Algebra in Relational Stream Systems

Data stream management systems make prominent use of a subset of relational algebra for describing continuous queries. The phenomenon has highlighted the wide spread use of declarative interface for stream data manipulation. As relational algebra is set oriented while the data streams are essentially sequences of data, there exist some cases where the use of relational algebra to describe processing in a data stream management system is counter productive. Supporting event algebraic operations on stream systems represents one such case. We have developed a combination of sequence algebra and relational algebra that is capable of supporting event algebra. Our alternative approach has been compared for performance and ease of specification. Unlike previous such attempts that exclusively focused on pattern detection in sequences of data, our work focuses on use of sequence operators to support a subset of event algebra. Additionally, event-condition-action (ECA) paradigm is used to create a Turing complete subset of extended SQL that supports relational and event algebra for data stream systems