Cooperation Scheme of Operating Systems that Merges Upper Layer Functions for Endhanced Security Infrastructure

信息系统正在成长为今天的社会基础设施，因此它的可靠性是我们健康生活的一个非常重要的因素。为此，人们对安全基础设施进行了广泛的研究，包括入侵检测系统、加密算法、认证方案、信息取证等新技术，但是，人们默认这些技术工作的基础层是可信的。这一基本层就是我们所说的操作系统，如果操作系统易受攻击，这些技术可能毫无用处。此外，信息系统通常被构造为分布式系统，但是这些系统的安全机制仍然被设计为以单主机系统的方式工作，并且缺乏作为多主机系统的视图。因此，我们明确了一些功能的要求，在操作系统作为一个社会基础设施，协同工作的多主机系统，并提出了一个基于策略的方案，在多主机系统的分布式访问控制。我们的建议有两个组成部分：一个是政策描述系统，可以分配足够的权力，每个主体的最低特权，另一个是机制，可以控制任何访问，以执行基于政策的信息流。

Information systems are growing up to be a social infrastructure today, so its dependability is a very important element for our sound lives. For this reason, much research is widely done for the secure infrastructure including new technologies in intrusion detection systems, encryption algorithms, authentication schemes, information forensics and so on ; however, it is tacitly assuming that the basic layer on which these technologies work can be trusted. This basic layer is what we call operating systems and these technologies can be useless if the operating systems are vulnerable. In addition, information systems are frequently constructed as distributed systems, but security mechanisms for these systems still are designed to work as a single-host system manner and lacks a view as a multiple-host system. So, we clarified some requirements for functions in terms of operating systems as a social infrastructure that work cooperatively as a multiple-host system, and proposed a policy-based scheme for distributed access control in multiple-host systems. Our proposal has two components : one is the policy-description system that can assign sufficient authorities for each principal of least privilege, the other is the mechanism that can control any access to enforce policy-based information flow.