Harmonic Proportional Bandwidth Allocation for Service Differentiation on Streaming Servers

Recent advance of real-time audio/video transcoding technology enables streaming servers to manage the constrained network-I/O bandwidth available to the Internet at a fine-grained level. In this article, we propose a bandwidth allocation mechanism for differentiated services (DiffServ) on the streaming servers. It aims to deliver high bit rate streams to high priority requests without overcompromising low priority requests. We formulate the bandwidth allocation problem as optimization of a harmonic utility function of the stream quality factors and derive the optimal streaming bit rates for requests of different classes under various server load conditions. We prove that the optimal allocation scheme, referred to harmonic proportional allocation, not only maximizes the system utility function, but also guarantees proportional sharing between classes with different pre-specified differentiation weights. We evaluate the harmonic allocation scheme via extensive simulations and compare it with an absolute DiffServ strategy and a proportionalshare strategy tailored from relative DiffServ in networking and operating systems. Simulation results show that the optimal allocation can meet the objective of relative DiffServ in both short and long timescales and greatly enhance the service availability and maintain low queueing delay when the streaming server load is high.