Anonymous Group Communication for Internet Services based on Publish/Subscribe

点击翻译按钮获取中文摘要

Many services on the Internet, like for instance Twitter, or WhatsApp, organize their participants into groups of users who share a common interest, and implement some group communication. Privacy in general, and specifically anonymous communication are increasingly important, in this context. This entails that while the content of some messages may be shared with large groups, the respective meta data of senders, receivers, as well as their interests must be kept secret. This privacy is expected towards 1) other members of the groups, 2) the service providers, and 3) eavesdroppers, like criminals or intelligence agencies of potentially totalitarian nation states. The initial negotiations in for instance dating or car-sharing sites may require that the participants can hide their identity, until some common trust has been established. Investigative journalism, or whistleblowing certainly requires the privacy of the users from service providers and state institutions. The proposed project will investigate anonymous group communication for the above stated applications. Instead of applying server-based approaches, we will research decentralized group communication following the Pub/Sub paradigm, as it allows for setting up multicast groups without any centralized entities.We will focus on three specific aspects: i) the development of algorithms and protocols for efficient pub/sub multicas, ii) the corresponding adversary models, and iii) privacy metrics, and the bounds on the possible privacy of users within such systems. Bringing together groups from Darmstadt and Dresden, we will interleave our respective methods and a) create formal models of user behaviour, adversaries, and the systems, b) develop user and load models as well as simulation models to evaluate the developed systems and the state of the art, and c) integrate the approach into existing friend-to-friend networks, to analyse it under realistic conditions.