软件定义网络中SVC视频层数和路径的联合优化

Joint Optimization of Path and Layers about Scalable Video Transmission in Software Defined Network

可伸缩视频编码技术是解决视频应用中网络异构性和终端多样性的有力工具,目前的工作大多集中在选择传输的码流层数或者优化单层数据的传输路径上.本文提出了一种可伸缩视频传输联合优化算法,能够对层数和传输路径进行联合优化,克服了只调整码流层数时效率低下的问题,同时改善了只优化传输路径时浪费带宽的现象.该算法首先根据0/1多背包规划模型建立问题模型,然后利用遗传算法进行求解,一次性决策出层数和传输路径.此外,算法还采用自回归积分滑动平均模型预测网络状态,将预测结果用于决策中.最后在Mininet平台进行了仿真实验,实验表明本文算法在对网络干扰较小的情况下,能够提供高质量的、具有服务质量保证的可伸缩视频传输服务.

Scalable Video Coding is a powerful solution to video application over heterogeneous networks and diversified end-users. Recently, works mostly concentrate on which layers to be transported or optimizing transmission path for single layer data. The emergence of Software Defined Network makes joint decision on choosing scalable video＇ s layers and planning each layer＇ s transmission path possible. This paper proposes an algorithm based on Genetic Algorithm for scalable video, which is a joint optimal layer selecting and routing. The algorithm uses 0/1 knapsack programming model to set up model, and predicts the network states by Autoregressive Integrated Moving Average Model, and gets decision based on Genetic Algorithm. At last it uses OpenFlow to transport streams ac- cording to the decision. Finally, experiments are carried out in Mininet to verify the feasibility and performance of our proposed algorithm. The experimental results show the proposed algorithm can provide high quality and quality-of-service supporting for scalable video transmission, while have little interference on network.