Producing synoptic data structure for high speed data streams in telecommunication network management

In large telecommunication network management system, substantial data containing the information of network traffic, network element status, device running situation and all other messages are continuously sent from each special network management system to the integrated network management system. This kind of data is typically the stream data. Current network management system employs traditional database technology to store and process such kind of data. However, the managed network grows larger and larger over time, and the data cumulates, too. The disadvantages of current technology gradually show up. Stream data usually arrives continuously and is often of high speed and large amount. Thus, a data stream algorithm based on self-adapting sliding window model is proposed here to build synoptic data structure in memory. By this algorithm, network manager is able to get the approximate query result after judging the accuracy requirement and processing ability.