CSR: Small: An Information Accountability Architecture for Distributed Enterprise Systems

个人身份或敏感信息（PII）已成为攻击者通过滥用其获取经济利益的目标。随着在复杂和不安全的系统上存储和处理PII的趋势，对改进保护的需求已成为企业政策和立法努力的目标。在这个项目中，我们研究了串联动态信息流跟踪（CDIFT），这是一种执行动态信息的架构。在分布式企业中对不同的系统级别和跨多个流程进行分析。CDIFT将允许管理员“映射”？企业业务逻辑（应用程序、网络、存储）并确定感兴趣的信息存储或传输的位置。同样的机制也可以用来强制执行信息。现在的政策，限制在哪里和谁可以查看这些信息。CDIFT将补充和加强当前的合规和审计工作，这需要管理员和审计员在了解现有系统方面进行大量的经常性工作和花费大量的工时。我们将开发和实验评估新技术，用于对感兴趣的信息进行细粒度跟踪（如系统操作员需要的，或者将来最终用户需要的，以灵活的、上下文敏感的方式），以映射这些信息通过企业的路径，并提供一种强制信息的方法。门禁和访问控制策略。我们的假设是，通过适当使用管理程序和分布式标记传播协议，可以创建有效的细粒度信息跟踪和访问控制机制，这些机制可以在整个企业遗留计算基础设施中运行。

Personally identifiable or sensitive information (PII) has become a target of attackers seeking financial gain through its misuse. With the trend toward storing and processing PII on complex and insecure systems, the need for improved protection has become a goal of enterprise policy and legislative efforts. In this project, we investigate Concatenated Dynamic Information Flow Tracking (CDIFT), an architecture for performing dynamic information flow analysis at various system levels and across multiple processes in a distributed enterprise. CDIFT will allow administrators to ?map? the enterprise business logic (applications, network, storage) and determine where information of interest is stored or transmitted. The same mechanism can also be used to enforce an information flow policy, restricting where and by whom such information can be viewed. CDIFT will complement and enhance current compliance and auditing efforts, which require considerable recurrent effort and a large number of man-hours spent by administrators and auditors on understanding existing systems.We will develop and experimentally evaluate novel techniques for conducting fine-grained tracking of information of interest (as defined by the system operator or, in the future, by end-users, in a flexible, context-sensitive manner) toward mapping the paths that such information takes through the enterprise and providing a means for enforcing information flow and access control policies. Our hypothesis is that it is possible to create efficient fine-grained information tracking and access control mechanisms that operate throughout an enterprise legacy computing infrastructure through appropriate use of hypervisors and distributed tag propagation protocols.