On the Impact of Dynamic Routing Metrics on a Geographic Protocol for WSNs

Abstract

Message routing is a keystone of Wireless Sensor Networks (WSN). In many routing algorithms, sensor nodes forward data in a multicast fashion and a specific routing metric defines which of the receiving nodes is the next forwarder. If this metric is static, optimal routes tend to be overused and nodes in that route die quickly. The Trustful Space-Time Protocol (TSTP) is a cross-layer protocol designed to deliver authenticated, encrypted, timed, and georeferenced messages containing SI-compliant data in a resource-efficient way. It defines a novel, data-centric paradigm for programming WSNs and the IoT. In this paper, we discuss and evaluate dynamic routing in TSTP, which can explore routing metrics naturally given by the protocol's cross-layer, geo-oriented design, such as distance to the destination and message TTL (Time-to-Live). We introduce the idea of spatial distortion to map virtually any metric into a time-to-transmit offset, and evaluate the impact of different combinations of metrics. We assess the spatial distortion metrics on a TSTP implementation over IEEE 802.15.4 on the OMNET++ simulator, comparing their impact on the network in terms of end-to-end delay, fairness index, and delivery ratio.