Crowd buzz: scalable audio communication for MMVEs using latency optimized hypercube gossiping

Audio communication for massively multiuser virtual environments (MMVEs) comprises the provision of hundreds of thousands of users with their personal, position dependent compositions of audio streams, in real-time, at low latency, using non-realtime components connected through a heterogeneous network with diversely limited bandwidths and varying delays. In lack of functioning comprehensive solutions, users resort to detached and limited conferencing tools. In this paper, we present a peer-to-peer approach with proven latency bounds, a suggested architecture, and best practices for its implementation. We test-drove our prototype using the OMNeT++ simulator in conjunction with the DELFOI delay model. First latency results indicate feasibility for domestic scenarios.