网元性能及业务数据特征精准分析方法

传统的网络性能分析以及业务数据分析，都是通过网管KPI 确定的，网络指标出现异常则需要根据优化人员经验进行问题定位。对问题定位的效率取决于技术人员的水平，同时网络业务数据特征的分析粒度也很粗，不能精细化、全面化、多样化地分析用户特征。随着统一DPI系统的引入，展示了用户会话过程中涉及的最细粒度的业务状态以及网络信道质量。为了更加准确地反映小区网络质量和用户真实业务体验，所研究的网元性能及业务数据分析方法采用数据切片和关联的方式，更加精细化地将体现网络质量的控制面信令XDR（X data recording，呼叫数据记录的扩展）数据和用户面业务XDR数据关联在一起分析。实现分应用、分业务拓展的用户感知分析，明确质差用户或场景等情况下XDR的数据特征，从而提升网络优化效率、提高用户感知。

Accurate analysis method of network element performance and service data characteristics

The performance of telecommunication network was usually analyzed by the KPI of network management system.If there was something wrong with the KPI,it needed the network optimization personnel to position problems.The efficiency of analyzing result depended on the quality of the network optimization personnel.Meanwhile,the analysis of data performance was rough estimated.As the unified deep packet inspection could show out the smaller granularity of the service status and channel quality,the analysis method of service data and node performance was discussed.By data slice and data association of control plane XDR and user plane XDR,the performance was analyzed with smaller granularity to show the quality of the network and the experience of users.So the user experience was showed out by different applies and services with different scenes,and the data characteristics of the XDR were more clearly.Finally,it improves the efficiency of network optimizing,the user feeling and the capability of data realization.