Unleashing the power of querying streaming data in a temporal database world: A relational algebra approach

Modern data-intensive applications have to manage huge quantities of streaming/relational data and need advanced query capabilities involving combinations of continuous queries (CQs) and one-time queries (OTQs) also requiring the verification of complex temporal conditions. In this paper, we go beyond the disjointed panorama of current approaches and adopt a new holistic approach to the integration of stream processing capabilities into the temporal database world based on the streaming table concept. To this end, we propose a full-fledged query interface composed of a TSQL2-like query language with an underlying algebraic framework. The algebraic framework, which is aimed at implementing the query interface on top of a working DBMS, is made up of: (a) the extended temporal algebra TA* supporting OTQs with an hybrid temporal semantics (sequenced and non-sequenced); (b) the continuous temporal algebra CTA that extends TA* with window expressions for CQ specification; (c) the translation of CTA expressions into TA* ones that can be executed by a traditional DBMS with an extended kernel.