实施大规模分布式可视化操纵系统的一个重要瓶颈在于节点问海量数据的传输。P2P Streaming技术逐渐成为解决该瓶颈的可行方案之一。主流P2P Streaming模型由于实时性上的缺陷并不适于此类系统。该文通过引入服务质量的概念并采用启发式的算法,设计一种改良的P2P Streaming模型。仿真实验结果表明,新模型在平均延迟、首包到达时间等指标上优于现有模型。
Massive data exchange among nodes is a bottleneck of deploying distributed visual steering systems. P2P Streaming becomes one of the solutions. Current P2P Streaming models are not suitable for such applications because they are short on real-time interactivity. Typical models and design improved model are analyzed by applying QoS concept and heuristic algorithm for P2P streaming. It is indicated by the result of simulations that new model is better than typical ones in metrics such as average delay and time to first package.