Protecting mobile agents communications in pervasive networks with a trusted distributed mediator for ID-based RSA

This paper proposes a new method of data authentication and encryption for distributed networks supporting mobile software agents. Software agents are a valuable instrument in wireless distributed monitoring networks, since they can be used to concentrate the monitoring effort in certain areas where an event is taking place. This way events can be tracked in a dynamic and efficient way. Mobile agents have to send messages to each other in order to coordinate their action, and those messages need to be secured by crypto credentials. But when the agents are moved over a wireless networks, how can the credentials be protected from sniffing by an attacker, besides layer II encryption? And if a rogue agent is injected in the network is it possible to limit the damages it can produce? The proposed approach bridges mediated RSA with the trusted platform modules (TPM) in order to provide an efficient and secure communication between agents. The communication is secured using ID-based cryptography while maintaing the compatibility with standard RSA and eliminating the mediator introduced by mRSA. We will show that this approach is convenient in terms of traffic overhead, perfectly applicable to existing TPM specifications and able to limit the damage that both an external and an internal attacker can produce to the network