Admission control in interference-coupled wireless data networks: A queuing theory-based network model

Abstract

Mobile traffic demand varies significantly in time and space. Hence, wireless radio resources in hotspot areas and at peak traffic hours may be scarce. Consequently, special attention has to be paid to effects induced by admission control, i. e., blocking of data requests by base stations in case of high utilization or overload. Moreover, rising traffic demand requires denser deployments and frequency reuse one. Due to the resulting inter-cell interference, the base stations' utilizations have to be considered mutually dependent, which affects the admission control performance. In this paper, we extend a flow level model for elastic traffic, which explicitly takes into account the dynamic mutual inter-cell interference among base stations, by admission control. The model presented allows computing exact values for the average base station resource utilization, flow throughputs, and blocking probabilities. To analyze large networks containing many cells, we extend two approximation techniques, a state aggregation and an average interference approach, and compare them with the exact solution. Both techniques require far less computational effort and show remarkable accuracy. We believe that the extended flow level model is a positive step towards a more accurate, flexible, and holistic framework for network analysis and planning, and self-organizing network techniques.