Mobile Agent Security Through Multi-Agent Protocols

University Of North Texas, Denton TX

Investigators

Abstract

An increasingly promising and widespread topic of research in distributed computing is the mobile agent paradigm: code which travels in an autonomous or semi-autonomous fashion between hosts, carrying computation state with the agent and performing computations on the remote hosts. The objective of this project is to explore the problem of protecting agents from malicious hosts using multi-agent protocols. Potential advantages of the multi-agent approach over previous approaches, such as using a trusted third party, include reducing assumptions of trust; addition of fault tolerance and protection from denial of service and collusion; and reduction of communication and space overhead by distributed data and circuit re-encryption. The specific methods employed in this project will be both theoretical and experimental. In the theoretical realm, the project explores advanced multi-agent protocol design, with the goal of achieving provable security in the aggregate even though individual hosts may be untrustworthy. The experimental part of the project involves setting up a small cluster to simulate a setting of remote hosts, which can then be instrumented to test and measure communication and computational overhead for sample applications. The amount of work currently being done on topics related to mobile agents makes it likely that agents will play an important role in future technologies. The very nature of mobile agent computation makes security a prime concern, so the impact of the work in this proposed project could be great.