A self-adaptive ALM architecture for P2P media streaming

Multimedia streaming in P2P networks where participants are involved in real-time multimedia sessions is a challenge. A key issue for the management of large-scale media streaming is the provision of application layer multicast. In existing works, scalability and QoS are the main critical issues and ALM optimization represents a real challenge when considering physical parameters or constraints such as delay, nodes capabilities or network density. In this paper, we propose a self-adaptive architecture for media streaming applications in a hybrid P2P network. This architecture supports dynamic and on demand scalability while limitating the overhead for multicast tree construction process. It adapts efficiently to network dynamicity according to a self and autonomous switching mechanism. The mechanism allows to switch from a centralized to a decentralized topology and vice versa, under some specific conditions. Simulation results show that our solution can be deployed at a large scale and gives satisfactory performances.