一种基于深度报文检测的合成视频质量评估方法

Quality assessment method for synthetic video based on deep packet inspection

移动视频业务快速发展,用户感知与网络KPI指标发生偏离,传统视频运维支撑手段存在端到端问题定界难的问题。本文对国内电信运营商移动视频业务特性进行分析,提出一种基于深度报文检测的视频合成算法,将事务级的DPI单据进行关联,以一个完整的视频业务为单位输出相关体验评估信息,进而研究以业务级为基础的DPI指标,用于评估用户观看完整视频的全过程感知,提升用户感知保障。此方法能够覆盖主流的视频网站资源,本文所提出的方法经过实际应用验证,可真实评估视频质量及用户感知体验,助力实现国内电信运营商通信网络问题定界与网络自动驾驶。

With the rapid development of mobile video services,user perception deviates from network KPI,and traditional video operation and maintenance support methods are diffi cult to demarcate end-to-end problems.This paper analyzes the characteristics of mobile video services of domestic telecom operators,and proposes a video synthesis algorithm based on deep packet inspection,which associates transaction-level DPI documents,and output relevant experience evaluation information based on a complete video service.Furthermore,it studies the DPI index based on the service level to evaluate the user's perception of the whole process of watching the complete video,and enhance the user's perception guarantee.This method can be applied to mainstream video website resources on the current network.According to theoretical analysis and big data collection from the existing network,it is proved that the proposed method can eff ectively improve the effi ciency of video experience analysis,realize end-to-end problem demarcation and active network operation and maintenance.