Spatial security policies for mobile agents in a Sentient Computing environment

Abstract

A Sentient Computing environment is one in which the system is able to perceive the state of the physical world and use this information to customise its behaviour. Mobile agents are a promising new programming methodology for building distributed applications with many advantages over traditional client-server designs. We believe that properly controlled mobile agents provide a good foundation on which to build Sentient applications. The aims of this work are threefold: (i) to provide a simple location-based mechanism for the creation of security policies to control mobile agents; (ii) to simplify the task of producing applications for a pervasive computing environment through the constrained use of mobile agents; and (iii) to demonstrate the applicability of recent theoretical work using ambients to model mobility. © Springer-Verlag Berlin Heidelberg 2003.