This paper explores the complications encountered when attempting to create a secure pervasive computing environment. The model introduced in this paper is primarily conceptual. The chosen environment is built around a shared space containing a classified network. Since access to the classified network is only available to people once they are in this space, we investigate what is required to ensure only authorised people can enter, which highlights problems inherent in any kind of physical access control, and then how to carry this through to manage their access to information once inside the space. A layered agent architecture is considered, using real time devices such as biometric mechanisms, door actuators and a smart card reader acting as the lower level agents. These agents are then wrapped in software allowing them to communicate with higher level agents that through the coordination of these low level agents control the security of the pervasive environment. Security is furth...
Patrick G. McLean