电信科学 ›› 2015, Vol. 31 ›› Issue (Z1): 249-253.doi: 10.11959/j.issn.1000-0801.2015399

• 其他 • 上一篇    下一篇

基于多层级联算法的网络业务流量识别技术

黄凤,王瑶,黄莉,姚继明,刘世栋   

  1. 国网智能电网研究院,江苏 南京210003
  • 出版日期:2015-12-20 发布日期:2017-07-03
  • 基金资助:
    国家电网公司科技项目“电力信息通信网络流量预测和管道智能化关键技术研究及其应用”

Traffic identification of network business based on multi-layer cascade algorithm

Feng HUANG,Yao WANG,Li HUANG,Jiming YAO,Shidong LIU   

  1. State Grid Smart Grid Research Institute,Nanjing 210003,China
  • Online:2015-12-20 Published:2017-07-03
  • Supported by:
    The Science and Technology Project of State Grid Corporation of China“Research and Application of Key Technologies for Traffic Prediction and Smart Pipeline in Power Communication Network”

摘要:

摘要:当前网络的主要问题是流量需求和网络带宽之间的矛盾,流量管理是有效的解决方案之一,同时能提升网络的服务质量,而网络业务流量识别是进行流量精细化管理的基础。在研究网络业务流量识别和机器学习技术的基础上,提出基于多层级联机器学习算法的网络业务流量识别方法,该方法能适应网络业务流量复杂多变的情况,并且平衡了机器学习算法在时间性能和识别准确率之间的矛盾,进一步地,该方法在级联算法上考虑了代价不均衡问题,更为关注重要业务的识别,有助于提升用户体验。

关键词: 流量识别, 智能管道, 多层级联算法, 机器学习

Abstract:

The contradiction between traffic demand and network bandwidth is the main problem of the current network.Traffic management is one of the effective solutions,at the same time it can improve the quality of service of the network,and network traffic identification is the basis of fine traffic management.The traffic identification method of network business based on multi-layer cascade algorithm was proposed,after the research of network traffic identification and machine learning techniques.This method can adapt to the situation of the complex network traffic flow,and can balance the contradiction between the time performance and the accuracy of the machine learning algorithm.Furthermore the cascade algorithm of this method takes the cost imbalance into account,and takes more attention to the identification of important business,so it can help enhance user experience.

Key words: traffic identification, smart pipeline, multi-layer cascade algorithm, machine learning

No Suggested Reading articles found!