一个基于TCR改进模型的网络流量管理系统①

网络流量管理技术是保证网络服务质量的重要前提。在研究网络流量识别与控制两种关键技术后，提出了一个实用的网络流量管理系统。在端口和特征值双层识别流量类型基础之上，针对TCP Ratecontrol（TCR）在往返时延的测量上难度大，精度低等不足，设计出了一种改进的流量控速模型，使得速率的控制更加有效和精确。最后在Linux的netfilter架构下实现了基于该模型的原型系统。通过实验测试验证了其有效性和准确性。

A Network Traffic Management System Based on Improved TCR Model

Network traffic management is a significant technology to ensure network QoS. At first, this paper analyzes the key technologies of network traffic management, including traffic identification technology and control technology. Then, a strategy of network traffic management system is proposed. To address the deficiencies of mensuration and precision of main network traffic rate control technology TCR, an improved traffic rate control model is proposed. This model makes the rate control more effective and accurate. Finally, a network traffic management prototype system is implemented based on Linux OS and netfilter firewall framework. Experimental test shows the effectiveness and the accuracy of this system.