Protected resource access for mobile agent-based distributed computing

This paper describes the mobile agent paradigm for distributed computing, and outlines the security-related issues encountered in supporting it. One of the major requirements is the provision of access control mechanisms for server resources. Several possible designs are discussed, and one such proxy-based design, which we have developed in conjunction with the Ajanta mobile agent system, is described in detail. Extensions of this mechanism which allow accounting of usage and selective revocation of access privileges are discussed. The Java security model is also described, since the security of the proxy mechanism depends on it.