Abstract

Early experiences in the WSN community, including our own, have reinforced the difficulties of developing large-scale applications. Extant analytical/formal models of sensor network hardware and simulation environments often lack the fidelity to adequately capture sensor and network behaviors of WSN applications when deployed in the field. As a result, the community has gravitated towards experimentation with current hardware and software platforms, not only to demonstrate applicability in real environments, but also to validate prototypes. Experimentation via at-scale field deployments has however been time-consuming and, in several respects, tedious. As a result, hardware and software testbeds are becoming a preferred basis for experimentation; they yield substantial efficiency in instrumenting potentially longlived experiments as compared to field deployments, which is valuable in the debugging, validation, and integration phases. They also enable research aimed at developing sound models and simulation tools, which if successful will in turn further expedite the development process. Some testbeds are sensing application- or domain-specific [1, 2, 3, 4]. Early adopter applications include security (of perimeters, buildings, etc.), monitoring (of agricultural lands, animal farms, forests, rivers, earthquakes, volcanoes, natural habitats, endangered species, factory

Citations