摘要： 提出一种基于查询请求分析的P2P网络访问热点的负载均衡方法。分析历史查询数据的时间序列，采用单指数平滑法预测未来的热点。根据访问热点的查询请求行为，采取端到端复制和轻负载邻居节点复制相结合的策略进行负载均衡，消除由访问热点造成的负载不均衡问题。实验结果验证了该方法的有效性。

关键词: 对等网络, 负载均衡, 查询请求, 时间序列, 热点预测, 单指数平滑法

Abstract: This paper proposes a P2P access hotspot load balance method based on query request analysis. A single-exponent smoothness method is adopted to predict the future hotspots through time series analysis of history query data, and then the combination strategy of end-to-end copy and light-load neighbor node copy is adopted to balance load according to the query request behaviors, which effectively prevent and eliminate the load imbalance caused by access hotspots. Experimental results validate the proposed method is effective.

Key words: P2P network, load balance, query request, time series, hotspot prediction, single-exponent smoothness method

中图分类号: 

• TN913.24