Towards Provable Transport Layer Privacy

Confidentiality of metadata is a challenging privacy problem. Most systems today ensure confidentiality of data using encryption, but they do not address confidentiality of metadata. For example, encrypted data still leaks metadata such as when a message is sent, how long it is, and to whom it is addressed. While metadata privacy is a known problem, existing solutions are far from perfect: they are either resource exhaustive which affects performance, or they guarantee only weak notions of privacy.

In this talk I present the problem of transport layer privacy, and outline a novel formalization of the problem using information flow control techniques that introduces a new trade-off between performance and privacy for anonymous communication. To exemplify I present a provably private protocol for instant messaging which we call Deniable Instant Messaging, DenIM for short.

Joint work in progress with E. Pagnin and A. Askarov.