Collaborative Research: Formal Privacy

点击翻译按钮获取中文摘要

Growing pressure to collect information on the Internet has created a need formore sophisticated ways to characterize privacy rights and balance them againstlegitimate commercial and law enforcement objectives. A wide range of businessesnow rely on their ability to collect information about online customers as afoundation of their added value in the market place. The use or potential use ofthe Internet for purposes such as child pornography and planning of hate crimesand terrorism have heighted interest in (and loosened regulations on) networkmonitoring. Theft of intellectual property has led to the development ofnumerous protection mechanisms; these often involve registrations and involuntarymonitoring of various kinds. In order to make security easier (and explore alucrative business model), a number of companies seek to act as third-partycaretakers of private information such as keys, authentication secrets, andcredit card numbers. Arrayed against these trends are many citizens incensed byaggressive means used to collect information from them and a variety of groupsthat champion privacy rights.There have been many advances in technologies both to aid information gatheringand limit it. One important trend is toward more advanced systems for creatingand managing digital credentials and authorization databases. In currentpractice credentials are sometimes `pushed' (like presenting a ticket to get intoa movie theater) sometimes `pulled' (like getting access to an airplane seat witha `paperless' ticket) and sometimes both (like making a purchase at a store witha credit card whose validity is confirmed online). These approaches haveInternet-based analogs, and technical advances have increased the range ofoptions considerably in recent years. For example, work on public key systemshas advanced techniques for delegation based on chains of `pushed' credentialsand increased the automation of credential collection. Another important trendis toward more advanced systems for protecting privacy using anonymizingtechniques. Tools such as onion routers and anonymous web publication serversprovide some support but other techniques directly aid fine-grained mechanismsfor obtaining privileges without exposing information unnecessarily. This work aims to develop formal support for characterizing privacy in thecontext of these advances. Efforts to improve access control systems andcredential distribution have paid little attention to privacy mechanisms so far,resulting in systems that are good at propagating credentials reliably, but nottuned to do so within well-understood privacy constraints. This collaborationwill build on our work in credential distribution and anonymity to create anintegrated architecture and protocols to provide advanced access control withinthe limitations of privacy constraints.