【作者】 郭杰；

【导师】 陈晓苏； 肖道举；

【作者基本信息】 华中科技大学 ， 计算机系统结构， 2004， 硕士

【摘要】 随着Internet技术的飞速发展,网络技术已深入到经济社会的各个层面。但是,由于Internet固有的开放性和匿名性特征,导致其安全问题成为了网络应用的主要障碍。在详细分析了安全套接层SSL(Security Socket Layer)协议实现机制的基础之上,研究了协议的安全机制,其中主要研究了SSL握手协议的安全机制。分析了协议实现时可能存在的三种安全隐患:证书发布时的安全隐患;证书撤销列表不能及时更新的安全隐患;在线证书状态协议不能提供实时更新的安全隐患。分析和总结了针对SSL协议的各种攻击方法,并对目前出现的一种针对SSL协议“认证服务器匿名客户端”模式的中间人攻击,分析了攻击过程以及攻击能够成功的前提条件。分析结果表明,如果客户端不能提供身份认证或者不能获得服务器证书的有效性检验,攻击者利用这种中间人攻击的方法就可以完全偷听会话内容。给出了防范中间人攻击的两种防范策略:随机数保护策略和客户端身份认证策略。引入这两种防范策略后,可对会话过程中的某些信息实行保护,保证了会话的安全。最后对防范策略从性能和安全性两个方面作了比较和分析。在分析的过程中,采用形式化方法描述了SSL握手协议的消息列表、各种攻击过程以及相应的防范策略的消息列表。分析结果表明,两种防范策略均能达到有效防止中间人攻击的目的。更多还原

【Abstract】 As Internet has been developing fast, network technology goes deep into almost every side of society and economy. However, security becomes the main obstacle of network application because of the inherent property of Internet for opening and anonymity.The research of the protocols’ security mechanism is based on thorough analysis of realization mechanism of every sub protocol of SSL. Security policy based on SSL handshake protocol was discussed. This paper analyzed the potential security problem may exist in realization of protocol such as those exist in certificate release, update process in canceling certificate lists and online certificate status protocol. Aiming at the newly appearing “man in-the-middle attack”, the attacking process and requirement for successful attack was analyzed. Analyzing result show that if the client can’t provide identification and can’t get the verification of the server certification, then attacker is able to use man-in-the-middle attack to eavesdrop the entire conversation. Two kind of defending strategy, which can be used to prevent man-in-the-middle attack effectively were discussed. By adopting this two strategies, certain information in conversation can be protected. At last, this paper analyzed and compared these two methods in term of performance and security. In analyzing the two methods, the message list of SSL handshake protocol was described, different attack process and message list of corresponding security policy using denotation defined by us.Experiment show that, these two methods can both prevents man-in-the-middle attack effectively.更多还原