摘要
在整个"软件人"安全体系结构中,授权是给特定的委托人授予特定资源访问权的一种访问控制机制,有必要对其内涵和表示形式进行深入探讨。扩展了大多数现有Agent平台所依赖的Java授权语义,总结出包括正向授权在内的5种授权机制,并且设计了适应"软件人"应用场景的新的权限类型。通过对比优选出适合"软件人"策略表示的方案,提出了一种具有高扩展性和灵活性的基于策略的"软件人"授权模型。
Authorization model is a mechanism for those given consigners to visit and control their special resources in whole SoftMan security architecture, so it is necessary to go deep into its connotation and denotation. The paper extended the Java＇s authorization semantics which has been popularly used in many Agent platforms,and there are five kinds of authorization mechanisms including positive authorization. A new permission types and a policy presentation schema were designed which is exactly suitable for SoftMan scenarios. A kind of policy-based SoftMan authorization model with high scalability and flexibility was put forward.
出处
《计算机科学》
CSCD
北大核心
2009年第2期275-277,286,共4页
Computer Science
基金
国家自然科学基金项目(60375038)
中国科学院自动化研究所复杂系统与智能科学重点实验室开放课题(20060105)
国家“十一五”科技支撑计划(2006BAJ04B07-2)
北京市自然科学基金项目(4072018)资助
关键词
人工生命
分布式系统
“软件人”
安全策略
授权模型
Artificial life, Distributed system, SoftMan, Security strategy, Authorization model
