摘要： 针对多媒体传感器网络中基于异常事件驱动的服务模式和基于用户查询驱动的服务模式，分析不同服务的流量模式，提出一种服务感知的公平调度策略，采用基于流量预测的动态公平缓冲共享机制和速率分配算法实现2类数据QoS的公平保障。仿真实验表明，该算法在传输延迟和成功到达率方面实现了公平性目标。

关键词: 多媒体传感器网络, 流量预测, 服务感知, 公平调度

Abstract: Aiming at two basic service models in Multimedia Sensor Networks(MSNs): event-driven and query-driven, this paper analyzes traffic patterns in various service models based on Differentiated Service(DiffServ). A Fairness Schedule Based Traffic Prediction(FSTP) is proposed, which achieves fairness based on the given QoS of two service flows. FSTP uses a Dynamic Fairness Buffer Sharing(DFBS) mechanism and rate allocation algorithm based on traffic prediction to ensure the fairness QoS between event-driven flows and continuous flows. Simulation results show that FSTP can achieve fairness in packet transmission delay and packet loss rate.

Key words: Multimedia Sensor Networks(MSNs), traffic prediction, service-aware, fairness schedule

中图分类号: 

• TP393