Privacy preserving database system

点击翻译按钮获取中文摘要

A Privacy Preserving Database Management System (DBMS) provides the core functions expected of any DBMS while guaranteeing data provided by a supplier is only used for its intended purpose. This research will develop a DBMS that includes privacy as a core feature rather than as an "add-on" to an existing non-privacy aware DBMS. The goal is to develop a system where data providers supply information with explicit privacy requirements, possibly different requirements for different kinds of data, while the DBMS itself can collect the data at any level but only use it in concordance with the privacy specified by the provider. Thus, the data provider's privacy is preserved by the DBMS.Once the privacy model is developed and demonstrated to be correct and complete, the approach is to take an existing DBMS, probably open source, and add these privacy features by wrapping an existing DBMS with them. This will not require changes to the DBMS engine itself and will provide insights into how to add privacy retroactively to legacy systems. Based on the lessons learned from this extension, the model will be incorporated directly into an open source DBMS engine to create a privacy-preserving one that is native in that all features are implemented within the engine itself. Thus, the native PP-DBMS will incorporate the privacy model in all components of a DBMS engine including the user interface, query processor, and transaction manager. This will ultimately lead to a system that can provide stronger security and provide users with increased privacy protection.The work will be undertaken in collaboration with several other academic researchers and industrial partners will be sought to develop commericialization possibilities.