Leaf Node Polling Model Analysis in an Integrated Access and Backhaul Network

In this paper, a polling service model is used for performance analysis of a leaf node in an Integrated Access and Backhaul (IAB) network. is constructed in order to analyse the performance measures of the system, which will enable the estimation of packet transmission delays in an IAB network. The Markovian polling queueing system with two queues and nonzero switching time is constructed to analyse the performance measures of the system, which will enable the estimation of packet transmission delays in an IAB network. One of the queues is designed to store packets transmitted in downlink from a parent node in the IAB network, and the second queue is designed to store packets transmitted in uplink from child nodes and user equipment associated with an analyzed node. Cyclic polling of queues allows to take into account the main feature of IAB technology — half–duplex mode of packet transmission. Using the apparatus of Markov processes, formulas for calculating the waiting time and the sojourn time of the request in the system, affecting the delay of packet transmission through the node, are obtained. The results of the numerical analysis illustrate an upper bound estimate for the server switching time at which the 5G NR delay constraints are fulfilled for networks with a single relay node. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.