Collusion Defender: Preserving Subscribers’ Privacy in Publish and Subscribe Systems

Abstract

The Publish and Subscribe (pub/sub) system is an \nestablished paradigm to disseminate the data from publishers \nto subscribers in a loosely coupled manner using a network \nof dedicated brokers. However, sensitive data could be exposed \nto malicious entities if brokers get compromised or hacked; or \neven worse, if brokers themselves are curious to learn about \nthe data. A viable mechanism to protect sensitive publications \nand subscriptions is to encrypt the data before it is disseminated \nthrough the brokers. State-of-the-art approaches allow brokers \nto perform encrypted matching without revealing publications \nand subscriptions. However, if malicious brokers collude with \nmalicious subscribers or publishers, they can learn the interests \nof innocent subscribers, even when the interests are encrypted.<br/> \n<br/> \nIn this article, we present a pub/sub system that ensures \nconfidentiality of publications and subscriptions in the presence \nof untrusted brokers. Furthermore, our solution resists collusion \nattacks between untrusted brokers and malicious subscribers (or \npublishers). Finally, we have implemented a prototype of our \nsolution to show its feasibility and efficiency.<br/> \n<br/><em>Index Terms:</em> Collusion Resistance, Secure Pub/sub, Subscribers’ \nPrivacy, Publications’ Confidentiality