• architectures
• multicores
• Parallel
• processing
• queries
• sliding-window

Data stream processing is an emerging computational paradigm to process unbounded flows of data received from sensors, IoT devices and from any other kinds of data sources. The thesis studies parallel processing techniques for sliding-window queries that are continuously processed over the most recent data. The work is focused on the categorization of different sliding-window models and the development of alternative models of a window-based parallel pattern. The thesis provides an extensive experimental evaluation that assesses the effectiveness of the proposed approaches.