将信息安全与系统自身业务及算法相结合,运用可信计算理论,提出了分层分布多Agent主动型通信安全防御框架。首先分析了网络攻击的特点和安全威胁的原因,给出了该防御框架的结构,研究其关键技术和策略,包括加密、签名、用户身份认证、访问权限与角色检查、平台认证、网络可信连接、可疑网络行为检查、加强型数字关联、与业务相关的容错处理和恶意行为分析等不同层次的可信模块。针对一种广域后备保护系统,给出了初步的安全设计,对变电站中多种加密算法的运行性能进行了测试。 Combining information security with the system’s own business and algorithms, this paper proposes a multi-agent multi-agent communication security defense framework based on trusted computing theory. Firstly, it analyzes the characteristics of network attacks and the causes of security threats. The structure of the defense framework is given. The key technologies and strategies are discussed, including encryption, signature, user authentication, access rights and role checking, platform authentication, Connectivity, suspicious network behavior checking, enhanced numeric correlation, business-related fault tolerance and malicious behavior analysis. Aiming at a wide area backup protection system, a preliminary safety design is given and the performance of various encryption algorithms in substations is tested.