摘要
恶意主机对移动Agent的威胁是移动Agent技术应用的主要障碍之一。对此,提出一种基于拆分-协作的保护移动Agent的方法。根据功能模块将一个移动Agent拆分成多个子Agent协同工作,完成原移动Agent相同的功能。借助信息论中的熵理论,证明了这种拆分使得移动Agent任务的理解难度随拆分数目的增长成正比增长,有助于保护移动Agent的机密性,从而在一定程度上阻止了恶意主机对移动Agent的攻击。最后讨论了效率问题及拆分的副作用。
In mobile agent application systems, security of mobile agents themselves is critical. According to the problems that malicious host may attack the mobile agents,we propose a method to protect mobile agents l^y splitting each one into a group of cooperative sub-agents which have the same function. Using Shannon＇s Entropy theory, we prove that the more fractions we split the agent into, the less likely one is able to understand it. This is helpful to protect the confidentiality and consequently prolong the attack time in some degree. The efficiency and side effect of the splitting are analyzed and the design and implementation issues of out prototype system are discussed.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第1期165-168,共4页
Computer Applications and Software
