Bandwidth prediction research based on BP neural networks in softswitch power dispatching network

Softswitch is a technology that the call controlling function are detached from the media gateway. As a new switch technology which is developing quickly, it has standardized interface, strong flexibility and open service. A reasonable allocation of bandwidth resources is necessary when building softswitch network. In softswitch network, the signaling stream is responsible for the transmission of signaling data which requires high security, reliability and low latency. So it seemed appropriate to make a prediction of the signaling bandwidth. By introducing the artificial neural network, the back-propagation(BP) neural network model can rationally solve the problem of prediction. The results show that when number of calls and average call duration per day were fed into the BP network the accuracy of bandwidth prediction is higher compared with it when the input factors were fed into the network individually. The BP neural network model can effectively predict the bandwidth of signaling stream in softswitch network and thus it well have some help to the bandwidth resources allocation when building a softswitch network.