Scalable Large-Scale Precise System-Wide Data-Driven Usage Control Across Layers of Abstraction and Across Machines

跨抽象层和跨机器的可扩展大规模精确系统范围数据驱动的使用控制

• 批准号: 183688753
• 负责人: Professor Dr. Alexander Pretschner
• 金额: --
• 依托单位: Lehrstuhl für Informatik XXII: Software Engineering
• 依托单位国家: 德国
• 项目类别: Priority Programmes
• 财政年份: 2010
• 资助国家: 德国
• 起止时间: 2009-12-31 至 2015-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/183688753?language=en
• 关键词: Scalable; Large; Scale; Precise; System

Usage control requirements stipulate constraints on the usage of data after access to them (“de-lete within three days,” “don’t copy.“) Control over data is, however, usually lost in distributed settings once the data is given away.The goal of the project is the run-time enforcement or at least the detection of violation of this kind of properties. Existing approaches focus on one concrete data representation, e.g., a file. In this project, a representation-independent solution is sought. To this end, existing concepts for reference monitors – that are usually defined for technical events – will, firstly, be extended by data flow detection: A deletion requirement for a file will then pertain to all copies of that file as well, the existence of which must hence be tracked. Secondly, a framework for the definition of precise technical machine-level semantics at different levels of abstraction will be provided: “copy” means, among other things, copy a file, copy&paste in Excel, and sending an email. To enforce data-driven usage control policies, reference monitors are, thirdly, defined at different levels of abstraction (e.g., operating system, runtime system, windowing system, separate IT system). Data flows will not only be monitored at each of these levels, but also in-between lev-els. For instance, under a strict interpretation of a prohibition to “copy,” the data’s path from a file through the operating system through the Java VM to native display functions must be tracked if one wants to, finally, prohibit copy&paste

使用控制要求规定了访问数据后对数据使用的限制（“在三天内删除”，“不要复制。“）然而，一旦数据被泄露，对数据的控制通常会在分布式设置中丢失。该项目的目标是运行时强制执行或至少检测对此类属性的违反。现有的方法集中于一种具体的数据表示，例如，一个文件。在这个项目中，寻求一个独立的解决方案。为此，引用监视器的现有概念-通常是为技术事件定义的-将首先通过数据流检测进行扩展：对文件的删除要求也将适用于该文件的所有副本，因此必须跟踪其存在。其次，将提供一个在不同抽象层次上定义精确的技术机器级语义的框架：“复制”意味着，除其他外，复制文件，在Excel中复制和粘贴，以及发送电子邮件。为了实施数据驱动的使用控制策略，第三，在不同的抽象级别（例如，操作系统、运行时系统、窗口系统、单独的IT系统）。数据流不仅会在每个级别上受到监控，还会在各个级别之间受到监控。例如，在对禁止“复制”的严格解释下，如果想要最终禁止复制和粘贴，则必须跟踪从文件通过操作系统通过Java VM到本机显示功能的数据路径