Investigation of an improved tensor routing model in the telecommunications network represented in the basis of interpolar paths and internal node pairs

Abstract

An improved tensor model of a telecommunication network (TCN) was selected for the study. The model is presented in terms of inter-polar paths and internal node pairs. The advantage of this model is the use of quality of service conditions in terms of bandwidth, average end-to-end delay, and packet loss probability. Within the framework of the chosen model, the technological routing problem is formulated in an optimization form together with the conditions and constraints imposed on routing variables. A system of criteria for optimality of routing solutions, which can be used to calculate optimal values for routing variables, is proposed. These criteria depending on the specifics of the routing problem focus on the optimal use of the available network resource and/or on the differentiated improvement of Quality of Service (QoS) indicators of network performance, to the numerical values of which one or another packet flow is sensitive. The analysis of the influence to the optimality criterion form on the nature of solutions to QoS-routing problems using the proposed tensor model of a telecommunications network is carried out. On a set of numerical examples, the improved tensor model of TCN both at the level of the proposed QoS-conditions and the considered QoS-optimality criteria confirmed its adequacy in solving the problems of a single path and multipath routing. The improved tensor model of TCN provided an adaptive response to changes in network load and the type of optimality criterion for routing solutions in terms of ensuring the required values of QoS-indicators – bandwidth, average end-to-end delay, and packet loss probability.