Application of convolution to estimation of cell loss rate in ATM networks

The dominant problems in ATM networks are network control and network resource allocation. An efficient resource allocation to each traffic source is an extremely complex task. Network must be managed in real-time in order to avoid congestion. Network congestion requires admission control that in turn requires traffic measurements. We propose a method for the estimation of a cell loss rate in an ATM switch that is based on the convolution of probability density functions of rate signals at the switch's input. The method is simple and can be easily implemented in a DSP in an ATM node. The results of cell loss rate estimation can be applied further to link capacity allocation and buffer design.