基于本体的多域访问控制策略集成研究-智能与分布计算实验室

基于本体的多域访问控制策略集成研究

姓名	赵战西
论文答辩日期	2006.05.08
论文提交日期	2006.05.15
论文级别	硕士
中文题名	基于本体的多域访问控制策略集成研究
英文题名	Research on Ontology-based Integration of Multi-domain Access Control Policies
导师1	李瑞轩
导师2
中文关键词	分布式系统;多域;访问控制;BLP模型;策略集成;本体;描述逻辑
英文关键词	Distributed system;Multi-domain;Access control;BLP model;Policy integration;Ontology;Description Logic
中文文摘	多域间的安全互操作，需要集成、协同和统一的安全管理。基于策略的安全管理是安全管理领域发展的里程碑，与多域分布式系统的安全管理的目标相适应，以安全策略为核心，构建集成环境；通过安全策略的规范表示和检验，保证安全策略与目标的统一。本体是共享概念模型的形式化规范说明，是一种能在语义和知识层次上描述信息系统的概念模型的建模工具，为解决多域安全互操作提供了一个新的方法。在单域环境中，针对经典的BLP（Bell La Padula）模型无法控制间接信息流的缺点，借鉴信息流模型的思想，在系统状态中增加记忆分量，记录主体所读的客体，控制信息向安全的方向流动。为了提高BLP模型的可用性并增强信息的完整性，将主体的安全级扩充为读写分离的安全级区间。通过这些方法将BLP模型改造成为具有记忆能力的新型多级安全模型。基于角色访问控制模型的一个重要特点是它是策略中性的，分析基于角色访问控制模型配置单域访问控制模型的方法，包括配置基于访问历史的BLP模型，使单域访问控制模型统一为基于角色访问控制模型。使用本体及其描述语言，对基于角色的访问控制策略进行了描述，形成一个概念和属性的公理集合（TBox），并采用ALCN（含有个数限制和补算子的描述逻辑语言）对TBox进行了形式化的描述。分析和改造IRBAC2000模型（一个基于角色转换的域间互操作模型），解决和避免该模型在多域集成中存在的一些问题。使用基于规则推理技术，定义多域访问控制模型中的一系列推理规则，实现访问控制领域的推理。最后，设计并实现了一个多域访问控制策略集成的模拟系统。在该系统中，单自治域内的访问控制策略定义在知识库的实例集合（ABox）中。系统采用单一本体方法，TBox、推理规则和ABox集成为全局共享的访问控制策略本体。
英文文摘	Multi-domain secure interoperation needs an integrated, cooperative and uniform secure management. The policy-based secure management is a landmark of the development in the secure management’s field and fits the goal of the secure management of multi-domain distributed system. The secure policy is the core of the policy-based secure management. And it can construct the integrated environment. It guarantees the policy fitting the goal by formalizing and checking the secure policy. Ontology is a formal, explicit specification of a shared conceptualization. It is a modeling tool that can describe the concept model of information system on the semantic and sciential level. It can provide a new way to solve the multi-domain secure interoperation problem. In the single domain environment, in order to control the indirect information flow, a History-based BLP model(HBLP) employs the idea of the information flow model, and adds the memorial factor in the system states to record the read objects and to control the direction of information flow. In order to improve the availability and the information integrity of the model, the HBLP model also extends the security level of the subjects into separated reading and writing security level ranges. By using these methods, we develop the BLP(Bell La Padula) Model into a new Multi-level Security Model with memory ability. An important characteristic of Role Based Access Control (RBAC) model is that it is policy neutral. We analyze the method how RBAC can be configured to enforce the single domain models, including the HBLP and unify the various access control models into one RBAC model. We use ontology and its description language to describe the access control policy, and create an axiom aggregation(TBox) to include the concepts and properties. We use the ALCN (a type of description logic includes the number restrictions and negation concepts) to formalize the TBox. We also analyze and improve the Secure Interoperability Using Dynamic Role Translation(IRBAC2000) model, to solve and avoid the problems that are happened in the multi-domain integrated environment. Through the Rule-based Reasoning Technology, we define the serial reasoning rules(Rule) of multi-domain access control model and enforce the rule in the multi-domain access control environment. Finally, a simulating system of the multi-domains access control policy is designed and implemented. In the system, the access control policies of the single domain are defined in the assertional(ABox). By using the single ontology method, the system integrates the TBox, Rule and ABox into the global access control policies ontology.