Stream Processing in Networks of Smart Devices

Abstract

Increasing attention on data streams has led to research in developing systems for querying streaming data. Special challenges arise, when data are gathered using networked devices which have constrained resources. Examples for such systems are sensor networks where motes can use their processing capability to reduce communication and save energy. Yet, it is still an open research issue how stream queries can optimally be processed in such networks. In this paper we address the problem of mapping stream queries to heterogeneous device-networks. We present a metric for handling the trade-off between quality of service (QoS) and energy consumption. This includes a model which describes how query operators and their hosting devices affect energy consumption and QoS. Based on the model we have developed an algorithm that optimizes mappings and the mapped query plans for a given network. We also discuss results calculated by a prototype that uses our metric