Abstract
The live broadcast business has developed rapidly in recent years, but low latency, smoothness, and high user quality of experience (QoE) are still the problems that we need to solve. At the same time, the arise of edge computing has made up for the lack of traditional cloud computing in live broadcast services and brought low-latency effects, but also brought a heavy burden to the edge network, making edge load balancing an important issue to be solved urgently. Meanwhile, the real-time and low-latency characteristics of live broadcast put forward higher requirements on the network, which increases the difficulty of load balancing optimization. Therefore, to cope with the development trend of adaptive live streaming and improve the QoE, this article first proposes an adaptive live streaming transmission framework under the edge network, which integrates edge service clusters, multicast, and client adaptive bitrate(ABR) algorithms. Then, we proposed a load optimization scheme for edge service clusters, focusing on the dynamics of load and transmission costs. According to the two factors directly affected by the ABR algorithm, i.e. edge server CPU load and bandwidth overhead, The dynamic overhead queue load-balancing algorithm is proposed. Compared with the other three load optimization algorithms, the simulation results show the superiority of our scheme in term of load balancing and overhead.