Towards Fault-Tolerant Mobile Agents

摘要

The absence of a trusted computing base for mobile agents poses serious security issues for both the host system and the survivability of the agent. Once a mobile agent is dispatched, asserting anything about the host system, the agent's behavior, or even the agent's existence is difficult to ascertain. In order to employ agents with any degree of confidence, constraints need to be placed on the agent computation since no restraints can be imposed (or assumed) about the host system's hardware or software. This paper presents a fault- tolerant approach for increasing an agent owner's confidence in the integrity of its agent.