Abstract
Modern multi-user communication systems, including popular instant messaging tools, social network platforms, and cooperative-work applications, offer flexible forms of communication and exchange of data. At any time point concurrent communication sessions involving different subsets of users can be invoked. The traditional tool for achieving security in a multi-party communication environment are group key exchange (GKE) protocols that provide participants with a secure group key for their subsequent communication. Yet, in communication scenarios where various user subsets may be involved in different sessions the deployment of classical GKE protocols has clear performance and scalability limitations as each new session should be preceded by a separate execution of the protocol. The motivation of this work is to study the possibility of designing more flexible GKE protocols allowing not only the computation of a group key for some initial set of users but also efficient derivation of independent secret keys for all potential subsets. In particular we improve and generalize the recently introduced GKE protocols enabling on-demand derivation of peer-to-peer keys (so called GKE+P protocols). We show how a group of users can agree on a secret group key while obtaining some additional information that they can use on-demand to efficiently compute independent secret keys for any possible subgroup. Our security analysis relies on the Gap Diffie-Hellman assumption and uses random oracles.