用户体验质量驱动的移动视频内容分发网络体系结构与关键技术研究

Now the network traffic is dominated by video, and most of user devices are connected to Internet in mobile mode. Under this trend, there are some great challenges for the current mobile Internet to provide video delivery service with the guarantee of user Experience of Quality (QoE), which include the architecture of video delivery, the video adaptation, and the cooperative management and control of diversified resource. Aimed at the core research goal, i.e. the QoE guarantee of mobile video delivery service, we plan to investigate the above key issues, and to obtain the following research achievements. The first is to establish an intelligent multi-layer multi-domain architecture of video delivery, which can promote the information exchange between user, video content, and wireless network, and guarantee the user QoE in cooperation mode. The second is to propose an adaptive coding theory and a streaming reconfiguration model for video content, which can improve the adaptability of video to conform the environment of network and device. The last is to establish a systematic technology of mobile video delivery integrated the cooperative allocation schemes for computing, bandwidth, and storage. Based on the above research activities, the innovations of the project consist of 1 architecture (i.e., an intelligent architecture of video delivery service), 3 mechanisms and models (i.e., video content reconfiguration, video adaptive coding, and Offloading-based cooperative delivery).

如何为用户提供具有体验质量QoE保障的移动视频分发服务，当前移动互联网面临着巨大挑战，主要表现在分发体系结构QoE考虑不足、视频自适应能力低、多维资源协同调控能力缺乏等三个方面。针对这些挑战，本项目拟围绕“移动视频业务的QoE保障”核心任务开展研究，其预期创新成果主要体现在：首先建立一种QoE驱动的多层多域移动视频智能分发体系结构，以促进内容、用户、网络三者信息交互、协同保障用户QoE；其次提出移动视频流自适应编码与快速重构理论与模型，提升视频内容对网络和终端的自适应能力；最后是建立计算、存储、通信等多维资源协同调控的移动视频内容分发技术体系，促进内容分发效率与资源利用率的最佳匹配。本项目的特色和创新主要体现在：1个体系结构(QoE驱动的智能分发体系结构)、3个机制和模型(视频内容快速重构机制、视频自适应编码机制、协同接力传输模型)。

在网络流量视频化和接入终端移动化的背景下，如何满足用户“随时、随地、随物”获得高质量视频业务的需求，当前移动互联网在分发体系结构、视频自适应能力、多维资源协同调控机理等三个方面面临着巨大挑战。本项目围绕“移动视频业务的QoE保障”核心任务，分别从这些方面开展了深入研究，并取得如下重要突破和创新：首先，以优化视频业务用户QoE为目标，基于Content-Centric和控制面与数据面分离思想，构建了基于SDN的多层多域移动视频智能分发体系结构，以促进视频内容、用户、网络等实体的信息共享和协同优化；其次，基于视频内容标识与用户体验意图紧耦合理念，构建一套面向内容信息特征的视频内容标识与解析机制，增强视频内容的全局易访问性和可用性；再其次，以中间媒体流(Meta-media)形式，建立面向异构网络和异构终端的视频流适配框架和视频流的快速重构模型，满足用户“随时、随地、随物”的视频访问需求；最后，从内容分发效率、资源利用率与用户QoE兼顾角度，建立了计算、存储、通信等多维资源协同调控的移动视频内容分发技术体系。这些研究成果能为我国内容分发网络的大规模部署及推广上提供理论依据和方法支持。本项目共获得技术发明专利8项、发表学术论文23篇(其中：期刊论文13篇、会议论文10篇)。此外，我们还成功主办了1次国际学术会议。同时，我们培养毕业博士生3人、硕士生15人。