Messaging in Sensor Networks: Addressing Wireless Communications and Application Diversity

In wireless sensor networks, which we refer to as sensornets hereafter, nodes coordinate with one another to perform tasks such as event detection and data collection. Since nodes are spatially distributed, message passing is the basic enabler of node coordination in sensornets. This chapter deals with the sensornet system service that is responsible for message passing; this service includes consideration of routing and transport issues and we refer to it as the messaging service. Even though messaging has been studied extensively in existing wired networks such as the Internet, it remains an important problem for sensornets. This is primarily due to the complex dynamics of wireless communication, to resource constraints, and to application diversity, as we explain below. Wireless communication is subject to the impact of a variety of factors such as fading, multi-path, environmental noise, and co-channel interference. As a result, wireless link properties (e.g., packet delivery rate) are dynamic and assume complex spatial and temporal patterns [41, 38]. The impact of these complex dynamics on messaging is substantial; for instance, it has led to changes in even the primitive concept of neighborhood [32]. Moreover, the mechanisms that deal with the dynamics are constrained in terms of their energy, network bandwidth, space, and time budget. Broadly speaking, the challenge then has been how to provide dependable, efficient, and scalable messaging despite the complex dynamics of wireless links. Sensornets have a broad range of application, in science (e.g., ecology and seismology), engineering (e.g., in-