Preventing Attacks on Mobile Agents by Malicious Hosts

The security concerns that come with using mobile agents is the main reason why the mobile agent paradigm has not been widely adopted. A number of new security problems are introduced with the use of mobile agents. Both, the host computer that executes an agent and the agent, need security mechanisms to protect against the threats posed by malicious agents and hosts. The issue of protecting a mobile agent from a malicious host is a more difficult problem than protect a host from a malicious agent. This paper will look at three solutions that have been proposed to protect agents from malicious hosts: creating time-limited black boxes, computing with encrypted functions, and environmental key generation.