Transformation of Pervasive Raw Data Into Actionable Knowledge∗

This paper presents an architecture for addressing the transformation of large volume of data generated by pervasive and other sensor applications in a timely manner (or subject to quality of service or QoS constraints). As the size and rate at which raw data is generated increases, novel approaches for meaningful reduction and extraction of actionable information from the data are needed. In this short paper, an architecture, based on the synthesis of stream and complex event processing (CEP) approaches, is presented. The proposed architecture is flexible and accommodates the use of a number of alternative or complementary techniques/approaches.