LayerP2P: A New Data Scheduling Approach for Layered Streaming in Heterogeneous Networks
Xin Xiao,Yuanchun Shi,Yuan Gao,Qian Zhang
DOI: https://doi.org/10.1109/INFCOM.2009.5061967
2009-01-01
Abstract:Although layered streaming in heterogeneous peer-to-peer networks has drawn great interest in recent years, there's still a lack of systematical studies on its data scheduling issue. In this paper, we propose a new scheduling approach for layered video streaming, called LayerP2P. The key idea and main contributions of LayerP2P come in two-fold: 1) According to the characteristics caused by layered coding, we propose four objectives that should be achieved by data scheduling: high throughput, high layer delivery ratio, low useless packets ratio, and low subscription jitter; 2) We design a 3-stage scheduling mechanism to request absent blocks, where the min-cost flow model, probability decision mechanism and multi-window remedy mechanism are employed in Free Stage, Decision Stage and Remedy Stage, respectively. Each stage has different scheduling objective while collaborates with each other, to achieve the above four objectives. Experimental results indicate that our approach outperforms other schemes in simulation environment. Besides, LayerP2P is implemented in the PDEPS Project in China, which is expected to be the first practical layered streaming system for education in peer-to-peer networks.