ABSTRACT
Beyond actual implementations using particular observational patterns and specific interpolation methods, a monitoring system can abstractly be described as a machine that generates knowledge about unobserved regions from a given set of observations. As such, it can be specified by its general structure, its resource requirements, and its interfaces towards data sources and higher level services like specific queries. This chapter proposes a general system architecture of a monitoring system as is induced by such an abstraction. It allows for an extensive analysis of the components of environmental monitoring regardless of the concrete implementation of a system. Only by such abstraction is it possible to compare the performance characteristics of different monitoring approaches on a sound common basis. The general properties of queries that are usually associated with such a monitoring are presented in this chapter.
