Exploiting cluster multicast for P2P streaming application in cellular system

Streaming over peer-to-peer (P2P) network is popular, however causes needless traffic traversal through multiple links due to the mismatch between the physical and the overlay network. Cellular channels are limited in number and expensive. Because of the magnitude of contents per unit time and the nature of playing same contents throughout the entire system, collaborative streaming approach is the key to an efficient P2P streaming system. In this paper, we propose a collaborative streaming system where some cellular peers download contents from the Internet peers, and then share the contents with the remaining cellular peers by employing device-to-device (D2D) multicast application in order to avoid bottleneck at the eNodeB (eNB), and to reduce streaming cost. We present the broadcasters/agents and their optimal assisted peers selection problem as stable admission assignment and formulate the problem as integer linear programming (ILP) problem. We also present a distributed algorithm to select agents among the cellular peers with suitable number of assisted peers for each agent to tackle the retransmission problem. The cellular peers change role as a broadcaster or a multicast receiver to ensure fairness. We also perform extensive simulation to show the efficacy of our design and to verify our claims.