摘要
移动代理作为一种分布计算和移动计算的支撑技术 ,得到广泛研究。但由于移动代理的迁移执行 ,引起一全新安全问题 :如何保护遭主机攻击的移动代理 ?即恶意主机问题 .从检测恶意主机攻击的角度出发 ,人们提出了基于参考状态 (reference states)的方法 .但目前所提出方法的开销都比较大 .认为只需记录移动代理在各主机运行时的输入和运行后的结果即可 .基于此 ,提出新的检测恶意主机攻击的协议 PTIR(a protocol tracing inputs and re-sults to detect m alicious host's attacks) .比较分析表明 。
As a basic technology of distributed computing and mobile computation, mobile agent becomes a focus of researches. Due to its migration in Internet and then execution in many unknown hosts, a new security problem is introduced: how to protect the mobile agent being attacked by some hosts? This is the problem of malicious hosts. In order to detect malicious host's attacks, a method based on reference states is suggested. But the overhead of this method is too large. In this paper, it is supposed to be enough to record the inputs and the results of mobile agent only. According to this point, a new protocol named PTIR (a protocol tracing inputs and results to detect malicious host's attacks) is built up to detect the attacks of malicous hosts. Compared with other protocols, the overhead of this protocol is low.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2002年第3期295-300,共6页
Journal of Computer Research and Development
基金
国家自然科学基金 (60 0 73 0 3 3 )
国家教委博士点基金资助