Protecting hosts against attacks in IMAGO system

A mobile agent is a piece of software which is able to migrate and execute on a remote host. The host may accept the agents without knowing the result in advance of executing the agents. Malicious agents may launch denial of service (DoS) attacks which may cause resource exhaustion or system deadlock. In fact, poorly written mobile agents or executing some agents in special cases can lead to the same result, so that authorized access of system services or resources does harm to the host as well. Mobile agents are coded in a variety of scripting or interpreted languages, and simply using semantic analysis for the predetection of potential hazards cannot provide a general solution. Discerning effectively, and thereby protecting the hosts against such potential threats, is necessary in mobile agent research. The paper presents a mobile agent system called the IMAGO (intelligent mobile agent gliding online) system and related security architecture. It introduces a way of protecting hosts against such possible threats. The paper demonstrates how the IMAGO system works against several typical threats with an insignificant overhead by the means of several techniques.