Analysis of mean delay in small cell networks with dynamic traffic and FIFO scheduling

Different network applications may have different delay requirements. How to evaluate the delay performance numerically is important for network design. Although the delay performances of random scheduling (RS) and round robin (RR) are known in literature, the accurate delay expression of first-input first-output (FIFO) scheduling remains to be discovered. In this paper, by combining stochastic geometry and queueing theory, we derive the accurate and tractable mean delay expression for FIFO, which accounts for the impact of random traffic arrivals and queuing interactions. The coupling relation between the success transmission probability and the network active probability is captured. Based on this, we further derive the mean packet delay of the whole network. The simulation verifies the accuracy of the analytical expressions. We find that for light traffic, it is best to choose FIFO from the delay aspect, and for heavy traffic, choosing RR is best. We also obtain the critical packet arrival rate for different scheduling selections.