Trust and Security for the Semantic Web

本文研究了分布式信任管理在动态和笔式计算环境中作为传统认证和访问控制方案的替代方案，如多代理系统、Web服务和普适计算。分布式信任管理利用人类社会中使用的技术来处理安全问题，在人类社会中，人们根据他们的能力、资产、关系和声誉进行评判。授权决策是通过应用相关的安全和信任策略来做出的，这些策略以高级声明语言表示。这些策略定义了代理(人或软件)的规则和约束，以及它们可以根据对象的凭据和属性对对象采取的操作。这项工作通过道义的权利、义务和禁止的概念扩展了信任管理的原则。基于语义语言(例如，RDF、DAML或OWL)的本体被用来表示构成凭证、策略、信念和证明的安全信息以及相关的领域特定属性，以表征代理、动作和对象。一个重要的结果将是更好地理解语义网如何在分布式开放环境中解决安全、可信隐私等关键问题。将通过实施原型工具和应用程序对新概念和新技术进行评估和演示

This research investigates distributed trust management as analternative to traditional authentication and access control schemesin dynamic and pen computing environments such as multiagent systems,web services and pervasive computing. Distributed trust managementhandles security with techniques used in human societies, where peopleare judged on their abilities, assets, relationships and reputations.Authorization decisions are made through the application of relevantsecurity and trust policies, expressed in a high-level declarativelanguage. The policies define rules and constraints on agents (humanor software) and the actions they can take on objects in terms oftheir credentials and properties. This work extends the principles oftrust management with deontic notions of rights, obligations, andprohibitions. An ontology grounded in a semantic language (e.g., RDF,DAML or OWL) is used to represent security information constitutingcredentials, policies, beliefs, and proofs as well as relevantdomain-specific properties to characterize the agents, actions andobjects. An important results will be a better understanding of howthe semantic web can address the critical issues of security, trustand privacy in distributed open environments. The new concepts andtechniques will be evaluated and demonstrated through theimplementation of prototype tools and applications