Message models and aggregation in knowledge based middleware for rich sensor systems

Networked, distributed real world sensing is an increasingly prominent topic in computing and has quickly expanded from resource constrained “sensor networks” measuring simple values to “sensor webs” of heterogenous networks encompassing many types of services and hosts, processing a wide variety of data and media. This paper presents ongoing work on OntoNet, which aims to provide messaging middleware in support of such rich sensor systems. In particular, this paper discusses the underlying message delivery model assumptions required in effectively supporting these settings. Those assumptions in turn present large implications for the mechanisms used to describe and match messages and destinations, as well as how to effectively do so in a scalable but correct manner. Initial concepts are also presented for two approaches to aggregating metadata and reducing network and memory consumption in OntoNet. One is a new application of least common subsumer induction, a known but infre...

• Type: Conference
• Year: 2009
• Where: DMSN
• Authors: Joseph B. Kopena, William C. Regli, Boon Thau Loo
• Keywords: DMSN 2009, Rich Sensor Systems, Sensor, Sensor Networks