Research of the properties of the monitoring model of hybrid networks of information systems

Abstract

To ensure reliable and high-quality service in information systems, it is critical to monitor and manage network load. This is especially relevant in the context of soft or hard real-time networks, where time delays can be critical. The expansion of the areas of application of wireless networks leads to the need to study the possibilities of their monitoring and management. When building hybrid networks, as a rule, the bottleneck is wireless segments, in which load assessment is difficult due to stochastic processes in the network. This paper presents a mathematical model for which a state graph of a queuing system is constructed for transmitting packets in a wireless network segment, a system of Kolmogorov equations is solved, and a distribution law is found. This model takes into account the modes of data transmission in the network and also allows for the identification of individual modes for deeper analysis. The paper shows that the best indicator of channel congestion is to determine the congestion based on the proportion of packets whose delivery time falls within the interval characterized by successful transmission after two unsuccessful attempts. The research results substantiate the methodology based on the allocation of time intervals for information delivery, which is based on the analysis of the principles of functioning and modeling of information processes. This method provides the opportunity to estimate network congestion with sufficient accuracy, based on experimentally obtained data delivery time. The methodology for assessing network congestion is based on modeling the ranges of information delivery times, which allows us to estimate the proportion of packets delivered in given time intervals. In addition, the model allows for a qualitative analysis of the contribution of packet delivery time probabilities for different transmission modes.