基于业务预测与聚类分析的网络疏忙评估方法

Network Load Alleviation Method Based on Traffic Prediction and Cluster Analysis

随着不限量套餐的推广,移动用户的业务结构发生了巨大变化,音乐、短视频等业务的流量冲击给运营商网络带来的压力亟待缓解。针对这一问题,提出了一种容量管理方案,运用业务预测方法,充分考虑实施提前量,预估网络业务量发展趋势;进一步根据预估业务量及网络配置情况划分业务场景,通过聚类分析,实施差异化的疏忙手段。工程实践表明,该方案能够有效缓解现网容量问题,大大提高了容量建设的精准性,并且具有良好的时效性。

With the promotion of unlimited packages,the traffic structure of mobile users has undergone tremendous changes.The traffic impact of music,short video and other services on the carrier network needs to be alleviated.To solve this issue,this paper proposes a capacity management solution through traffic prediction method,taking a full consideration of implementation amount in advance and estimating the development trend of network traffic volume.Furthermore,the traffic scenarios are divided by estimating traffic volume and network configurations,and differentiated alleviation methods are implemented via clustering analysis.The engineering practice shows that the proposed solution can effectively alleviate the existing network capacity problem,significantly improves the accuracy of capacity construction,and has good timeliness.