A sampling method for predicting end-to-end delay performance in wireless sensor networks

Abstract

End-to-end delay performance is an important Quality of Service (QoS) metric in 5G communication systems and wireless sensor networks (WSNs). Recently, a multi-hop effective capacity model was proposed to provide accurate characterization of end-to-end delay performance in wireless multi-hop environments. However, this model was developed in the continuous-time domain, which accounts for a discrepancy in digital/discrete-time systems. In this paper, we extend such a model into the discrete-time domain and derive new mathematical formulae for tail probabilities of delay, delay mean and jitter in multi-hop cases. Furthermore, we propose a simple algorithm for end-to-end delay performance prediction based on the sampling method. By using publicly-available real traces from a wireless sensor network, we recreate these field experiments in a simulation platform to validate the algorithm. The results show that the algorithm gives satisfactory prediction.