A tensor data-flow model of multipath routing with guarantee of quality by several indices is presented. The novelty of the model consists in transfer to the updated version of the flow storage conditions taking into account packets possible losses caused by overflow of the buffer queue on the network routers. Due to the tensor formalization of the network model it was possible to receive in the analytical form the conditions of imparting quality to simultaneous service by multiple various indices, namely, the transmission rate, mean delay and packages loss probability. The results of the research have shown the advantages of the offered method; they have made themselves evident in the improvements in the end-to end multipath delay, increase in the transmission rate, packets jitter minimization resulted from realization of the routing multipath strategy, increase in the probability of meeting the QoS-requirements and decrease in service abandonments caused by nonfulfillment of the specified quality level.
Lemeshko A.V., Evseeva O.Yu., Garkusha S.V. Research on Tensor Model of Multipath Routing in Telecommunication Network with Support of Service Quality by Greate Number of Indices // Telecommunications and RadioEngineering, Vol.73, No 15. - P. 1339-1360.