A scheme of aggregating stream data

This paper proposes a scheme of aggregating tuple stream data. Tuples are handled as multidimensional data and the scheme aggregates fact data on the selected dimensions of the received tuples. A tuple server node directly gets tuples from a tuple generation source and sends them out to client nodes located on network. A client node filters the received tuples to construct its own fact table by using a multidimensional extendible array. The fact data associated with the filtered tuples are aggregated in the extendible fact table and used for multidimensional analysis (MOLAP). Such an extendible fact table is organized by an implementation scheme for multidimensional datasets based on history-offset encoding scheme. Different from the usual method employing fixed size multidimensional array, the proposed scheme can organize an extendible fact data table with sufficiently small storage just fits to the quantity of the received tuples.