ASSESSMENT OF THE STRUCTURAL RELIABILITY OF TELECOMMUNICATION NETWORKS OF UNCERTAIN TOPOLOGY BASED ON IMITATION MODELING

Abstract

This article is devoted to the analysis of existing and the development of new approaches to assess the structural reliability of networks with an uncertain topology. Ensuring reliability in modern telecommunication networks is becoming an increasingly important issue in connection with the intensive development of these networks over the past decade. Existing structural methods for evaluating reliability are focused on applications for networks with a known structure. However, in cases where the network topology is constantly changing, these methods are of little use. Therefore, an approach is proposed for obtaining estimates of the structural reliability of telecommunication networks with an uncertain topology based on the use of basic structural characteristics alone. The developed method for assessing the reliability of telecommunication networks with an uncertain topology is presented, which provides the ability to determine: the total number of routes in the network with a certain number of nodes and branches; the number of routes per one connection, as well as the number of routes of rank no more than the allowable value that can be used for one connection. The upper limit of the structural reliability of individual connections are determined by the routes obtained for each connection), on the basis of which the structural reliability of a network with an uncertain topology is evaluated. A simulation computer model has been created for assessing the structural reliability of a telecommunication network with an uncertain topology, which consists of a computer model that empirically estimates the number of routes in various implementations of obtaining indicators of the structural reliability of communication between a pair of nodes and the entire network as a whole. A simulation experiment of computer modeling with given structural characteristics consists of a series of tests, in each of which a specific random implementation of a network graph is generated and all the routes existing in it are searched. The article presents the main functions and components of the developed software system, the figures show the modules and the basic algorithms of the model. Interfaces for inputting initial data and for outputting simulation results in the form of values ​​of the number of routes based on the results of simulation experiments and corresponding graphs are demonstrated. A separate important issue that is solved in the simulation is the determination of a sufficient number of tests in an experiment. The criterion for assessing this modeling parameter was the deviation dynamics of the average total number of routes of ranks 1 through 4 obtained in the tests.