Study of multicast video streaming in cloud computing environment

Data center network architectures are widely applied in cloud computing. Clients requesting access to multiple servers in multicast environments typically results in throughput collapse because of transmission control protocol (TCP) retransmissions and limited switch buffer sizes. This communication pattern is known as TCP Incast. This study investigated the network performance of a cloud environment with multiple users receiving live TV streaming. The digital fountain concept was adopted as a network transmission model and the user datagram protocol was adopted instead of TCP. Moreover, Luby transform codes were employed to maintain data accuracy and scalable video coding (SVC) was employed for video compression, in which streams were divided into base layers and enhancement layers to provide varying levels of service. Each video frame has a time-validity attribute. The proposed digital fountain based SVC multicast method was designed to enhance network performance and improve quality of service. Finally, a simulation was performed using a local area network architecture to compare the outcomes under various operating conditions. The simulation results show that enhancement layer was streamed regardless of the reception of the base layer - even no one was receiving the enhancement layer under a low-bandwidth network.