| This thesis develops procedures to increase the efficiency, security, and reliability of software agent communications. Software agents are special types of software programs that can migrate through a network. They are autonomous and can be intelligent. Software agents can search for and interact with services to perform tasks on the user's behalf. Software Agents have become a promising technology, and they have been applied to a number of different areas. However, software agent communications are suffering from security hazards, network overheads, and the lack of reliability.; This thesis describes a security mechanism and a multiplexing technique for the TEEMA agent execution environment. The security mechanism provides secure channels for TEEMA, thus enhancing software agents' communications with user authentication and data integrity. The multiplexing technique reduces network overheads caused by processes establishing new channels.; A set of tests was performed to evaluate the quality and the performance of the agent platform. The security and efficiency were measured in terms of the time spent for tested agents to move repeatedly between the two agent platforms. Results indicate that a secure and reliable agent platform has been built and overheads have been reduced. |
