Processing algorithms for multi-priority traffic in switching nodes and their impact on the communication service quality

The article examines the issue of the service quality of multi-priority network traffic in multiservice military communications networks. A specific feature of military communication networks is the difference in the significance of the transmitted information, that is, the need to categorize it, and the explicit hierarchy of command units among which it is transmitted. This feature is taken into account when obtaining a composite quality measure, which is formed by additive folding of partial quality measures and weight coefficients, which are determined depending on the traffic category and type. Partial quality measures are formed from the values of the delay time and the probability of timely delivery of network traffic. The values of the timeout parameters and the probability of timely delivery of network traffic given in the article are obtained using simulation method. The simulation model of the switching node in the military communication network enables to obtain the value of the above parameters using various processing algorithms for queuing, namely: FIFO – first in, first out; PQ – priority queue; WFQ – weighted fair queuing. The simulation resulted in getting dependences of the delay time and probability of IP packets timely delivery on the use factor of switching node when using the above queuing disciplines. The histogram was constructed based on the obtained results as an example, after their normalization and getting the integral estimate of the partial quality indicator for each type of traffic and user group. This histogram reveals the impact of the queuing service algorithms on the service quality of various classes and groups of traffic with the preset use factor of the switching node. The composite quality measure is calculated. And the article presents the dependence of the composite quality measure for multi-priority traffic on the use factor of the switching node when using the proposed queue processing algorithms. Concluding the article, the authors assess the applicability of various queuing service algorithms with different use factors of the switching node.