基于Sketch的网络行为测量关键技术与系统

网络行为测量是对网络流量的特征和变化趋势进行测量和分析，可以解决网络运营、容量规划、网络账户计费、拥塞控制、流量调度、流量异常检测等问题。本项目基于Sketch算法研究两类网络行为测量：流量特征行为测量和网络异常行为测量。首先，为了完善网络行为测量的基础理论，本项目研究Sketch算法的误差理论和无偏理论。其次，本项目针对软件定义网络和时间敏感网络研究网络行为测量新方法。再次，本项目研究适合不同未来网络环境的流量特征及其提取方法。最后，本项目研究基于意图的网络行为测量技术，通过预定义的系统接口表达测量意图，并操作相关网络设备与数据实现意图。

Network behavior measurement is to measure and analyze the features and the trend of network traffic. It is able to solve problems such as network operation, capacity planning, network account billing, congestion control, traffic scheduling, and traffic anomaly detection. Based on Sketch algorithms, this project studies two kinds of network behavior measurement, behavior measurement on the features of the traffic and the anomaly of network. First, this project studies the error theory and the unbiased theory of Sketch algorithms to perfect the basic theories of network behavior measurement. Second, this project studies new methods of network behavior measurement for Software-Defined Networking and Time-Sensitive Networking. Third, this project studies traffic features and related feature extraction methods applied in different future network environment. Last, this project studies the technology of intent-based network behavior measurement. This project expresses measurement intent through predefined system interfaces and operates relevant network equipment and data to achieve the intent.