面向流媒体应用的实时可执行概率获取方法研究

Research on acquire technology of real-time executable probability facing streaming media application

为准确获取网络实时可执行概率,从而对网络实时任务进行接纳控制,是解决当前网络尽力服务与实时任务服务质量（QoS）要求间矛盾的重要手段。面向流媒体应用,针对网络资源余量的高度不确定性,分析和实时获取了变化中的网络资源余量的概率时效图,并对今后的时效曲线进行预测,得到流媒体应用的实时可执行概率。实验比较了资源余量的概率时效图的预测值和实测值,结果较符合,为共享资源网络上执行实时任务作出了有益的探索。

It is an important means to resolve the contradiction between the best-effort service of network and the quality of service for real-time tasks that the real-time executable probability is accurately acquired for real-time tasks’admission control on shared resource network. Facing the streaming media application,this paper analyzed the varied status of network resource in connection with the uncertainty of the surplus. And got the time-effectiveness curve diagram of the network resource margin and forecast the farther trend,so that it could obtain the real-time executable probability in the streaming media apllication. Finally,the results of experiment which compared forecasting value with real value show that the method is effective. This paper shows a beneficial study on key issues on real-time tasks executed in shared resource network.