Scaling Issues for Energy and Communication in Large Networks of Small Sensors

Interesting and important optimization problems arise in systems that are simultaneously very large and very small, for example, very large networks of very small sensors. Regarding energy and information, smaller nodes carry less onboard energy than larger nodes, and they are worse antennas for communication at longer wavelength, i.e., lower energy per photon, so they have smaller total information communication capacity. Regarding energy and lifetime, i.e., ratio of stored energy to baseline power demand, a node's baseline power demand usually scales as a smaller power of its linear dimension than does its energy storage capacity, so its operating time -and, if it is mobile, its operating range -usually falls more rapidly than linearly with smaller size. Depending on operating scenarios -e.g., whether or not you expect a mobile entity's speed to be independent of its size -and constraints - e.g., whether the system as a whole can use only a limited quantity of some critical material -tradeoffs arise between node size and node spacing. Node spacing affects both the power required to communicate between nodes, and the spatial density of gathered environmental data, often the essential measure of the deployed network's effectiveness. This paper demonstrates a general approach to analyzing -hence optimizing in the face of specified constraints -the competing consequences of large and small size, and examines them briefly but explicitly in plausible practical implementations.