基于拆分-协作方法的移动Agent安全保护  

PROTECTING MOBILE AGENT AGAINST MALICIOUS HOST ATTACKS USING SPLITTING AND COOPERATION

作　　者：宋亚奇[1] 房鼎益[1] 吴晓南[1] 陈晓江[1] 

机构地区：[1]西北大学计算机科学系,陕西西安710069

出　　处：《计算机应用与软件》2007年第1期165-168,共4页Computer Applications and Software

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.

关 键 词：移动AGENT 拆分 协作 恶意主机 安全 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]