User-centric, Secure Information Flow Management in Enterprise Systems (USIFES)

网络化企业信息系统的复杂性日益威胁着数据安全。传统的访问控制和授权方法之所以失败，是因为威胁通常是由授权用户无意或有意的活动引起的。在这个项目中，我们的重点是在系统和用户之间的接口控制不希望的信息流。通常不受欢迎的信息流不是来自恶意攻击，而是来自几个使用事件和信息交换的交互，而这些事件和信息交换本身似乎是无害的。但是，它们一起破坏了系统的安全策略。这个项目的目标是开发一种技术来识别、建模和防止由人机、人机和计算机-计算机交互之间的相互作用产生的不良信息流。为此，我们形式化了可能导致信息流的系统和用户操作，并开发了特定于领域的语言来指定此类操作的序列。我们通过(1)从攻击者的角度来识别这些行为，(2)从真实用户中挖掘信息流活动。我们的方法包括测试用例的定义，以验证、度量和证明系统对已识别的威胁的安全程度。我们通过在现有标准企业软件的环境中开发一个演示程序来验证我们的方法。

Data security is increasingly threatened by the complexity of networked enterprise information systems. Classical methods of access control and authorization fail because threats often arise from unintentional or intentional activities of authorized users. In this project, we focus on controlling undesired information flow at the interface between system and user. Often undesired information flows result not from malicious attacks but from the interaction of several usage events and information exchanges that, on their own, appear to be quite harmless. Together, however, they break the security policy of the system. The goal of this project is to develop a technique to identify, model, and protect against undesirable information flows resulting from the interplay between human-computer, human-human and computer-computer interactions. To this end, we formalize system and user actions that could potentially result in information flow and develop a domain specific language to specify sequences of such actions. We identify these actions by (1) taking the viewpoint of an attacker, and (2) mining information flow activities from real users. Our approach include the definition of test cases to verify, measure and certify the level to which system are secure against the identified threats. We validate our approach by developing a demonstrator in the context of existing standard enterprise software.