End-to-End Delay Analysis of a Wireless Sensor Network Using Stochastic Network Calculus

Understanding delay performance is essential for successful development and deployment of real-time networked sensing applications. In this work, we characterize the end-to-end delay on a multi-hop Wireless Sensor Network (WSN) with an analytical method based on Stochastic Network Calculus (SNC) with Moment Generating Functions (MGF). The particularity of the sensing scenario is that all nodes generate and cooperatively forward their traffic through the network to a main location, known as the sink. We tackled the problem by modeling the service process of fading wireless channels from a high-layer perspective with Finite-State Markov Chains (FSMC). Numerical performance bounds are provided for an example WSN with IEEE 802.11g ad-hoc links in which the effects of delay bound violation probability, per-node offered load, data rate and fading speed are quantified for different network sizes. Finally, the presented analysis is validated through a comparison between analytical and numerical simulation results.