Distributed Security in Multi-Agent Systems

The use of mobile agents in distributed computing represents an alternative to the conventional client-server model. Although agents are present in many critical applications, the security problem in such systems has not been sufficiently studied yet. This paper presents a new method for assuring security in a multi-agent system by using an extension of the ElGamal encryption algorithm. The messages sent between the system's components can be decrypted only if the agents that hold the parts of the private key work together.