Publications‎ > ‎

Distributed Soft Security for Mobile Agents


A mobile agent is an autonomous agent that is not bound to the system on which it begins execution; it is created in one execution environment and can travel to a different host, carrying both its state and its code. Unfortunately, the mobility and interaction inherent in mobile agent technology introduces significant security problems. Agents may relocate to a host environment in which the host acts maliciously, disabling the agent, altering its code, or disrupting its communication. In addition, altruism among agents cannot be assumed in open environments; mobile agents may come in contact with other agents that can disrupt agent activity or misrepresent information. The Buddy Model is innovative because it provides distributed security based on cooperating agents that organize to efficiently protect each other from malicious servers. The Buddy Model stops short of implementing an actual rescue algorithm, however. This research extends the Buddy Model by overlaying a strategy by which mobile agents assist each other in avoiding malicious hosts and recovering from host attacks. This extension relies on the notion of “soft security,” by which agents model the trustworthiness of hosts. Experiments show that the incorporation of soft security to the Buddy Model allows mobile agents to better choose trustworthy Buddies and avoid attacks by malicious hosts.

D. DeAngelis and K. K. Fullam. Distributed Soft Security for Mobile Agents. Technical Report. TR2005-UT-LIPS-003, The University of Texas at Austin, Austin, TX, 2005.

Dave DeAngelis,
Jun 4, 2009, 1:55 PM