Processing of Data Streams with Prediction Functions

Abstract

Networks of sensors arise naturally in many different fields, from industrial applications (e.g., monitoring of environmental parameters in a chemical plant) to surveillance applications (e.g., sensors that detect the presence of intruders in a private property). The common feature of these applications is the necessity of a monitoring infrastructure that analyzes continuous supplies of data streams and outputs the values that satisfy certain constraints. In this paper we present an approach to process monitoring queries in a network of sensors with prediction functions. We consider sensors that communicate their values according to a threshold policy and our query processing leverages prediction functions to compare tuples efficiently and generate answers even in the absence of new incoming tuples. We deal with two types of constraints: window-join constraints and value constraints.