数据驱动的网络信息平面研究

互联网承载的不断增长且日益多样化的信息业务需求与网络的性能和灵活性不足之间的矛盾日益严重，而网络和应用之间缺乏相互感知的能力是导致该矛盾的一个重要原因。项目提出了使网络和应用解耦、独立的数据驱动的信息平面模型，通过多维多源网络参数测量与异构多样化应用测量两类关键技术实现信息平面构建，以提高网络与应用之间双向感知的精度、效率和灵活性。同时，项目基于信息平面开展数据驱动的网络资源管理与任务调度优化、以及网络体系架构的演化研究，以进一步展示数据驱动信息平面在网络体系结构优化中的作用。项目拟发表高水平论文15篇左右，其中至少2篇CCF A类会议或期刊论文、拟申请发明专利3-5项。

Increasing requirements on various Internet services cause severe challenges to the service performance and network flexibility. However, there still lacks a mutually-sensing capability between the Internet and its services, which has become an important factor in causing such challenges. In this project, we propose an independent data-driven information plane (DDIP) for Internet services, which bridges Internet and services in a reliable, flexible and efficient manner. Meanwhile, we propose to improve the network resource deployment and task scheduling as well as the study on Internet architecture based on DDIP to show the advantages of it. We aim to publish around 15 papers with at least 2CCF A-class papers, and to file 3-5 patents during the project.