Multistage Congestion Games for live streaming

We model peer-to-peer live streaming as a multistage congestion game where certain strategy restrictions allow, at equilibrium, to minimize both streaming duration and congestion. We also propose a distributed algorithm (ConGaS) that can be easily executed at peers, enabling them to coordinate toward streaming optimization. Finally, ConGaS is compared against two other dissemination policies through experimental evaluation, and simulations confirm the viability and efficacy of the former.