CT-ER : Runtime Techniques for protecting confidential data in large scale software

随着计算机软件处理的机密信息越来越多，确保个人信息的隐私成为人们非常关注的问题。为了解决这一问题，本项目研究了跟踪软件程序处理的敏感数据使用的机制，目的是防止未经授权的泄露。技术重点是使用跟踪敏感数据使用的运行时机制。本提案研究了三个重点领域：(a)基于流程的分离机制和执行保密策略的运行时检查机制；(b)在计划使用后从程序中删除敏感数据的技术；(c)跟踪隐含信息流的技术。该项目使用了一些新颖的想法，这些想法建立在PI先前在基于语言的技术方面的工作上，以解决这些重点领域的基础研究问题。为了最大限度地发挥影响，该项目还将实现将这些技术应用于大规模软件应用程序的工具。这些工具将使程序员和最终用户能够改进程序，以保护他们的机密数据。这些工具将通过互联网分发，以尽可能广泛地传播和进一步改进。在教育方面，本项目旨在培养研究生和本科生在基于语言的计算机安全和隐私方法应用方面进行研究。

With the increasing amount of confidential information being handledby computer software, ensuring privacy of personal information isbecoming a great concern. To address this concern, this projectinvestigates mechanisms that trackthe use of sensitive data handled by software programs with the goal ofpreventing unauthorized disclosure.The technical focus is on the use of runtime mechanisms that track theuse of sensitive data. There are three thrust areas investigated inthis proposal (a) mechanisms for process based separation andruntime checking for enforcing confidentiality policies(b) techniques for erasing sensitive data from programsafter their intended use and (c) techniques for tracking implicitinformation flows. The project uses novel ideas that build on the PI'sprior work in language based techniques for addressingfundamental research questions in these thrust areas.To maximize impact, this project will also implement toolsthat will apply these techniques to large scale softwareapplications. These tools will enable programmers and end usersto retrofit programs in order to safeguard theirconfidential data. These tools will be distributed using theWeb for the widest possible dissemination and further enhancement.With regards to education, this project aims to train both graduate andundergraduate students to perform research in the the application oflanguage based methods for computer security and privacy.