摘要
针对Skyline计算中,需要处理的数据量大,处理时间较长的问题,引入P2P网络,将数据计算的压力分摊至各网络节点.预处理中,合理采用数据映射方式,增加同一节点数据间的决定能力,减少本地计算量.在全局Skyline计算时,通过网络点对点传输,将各节点需计算数据量减少至最小.实验结果和理论分析表明,新算法可将Chord网络中,本地节点需要计算的数据量减至10%左右,当数据量较大,数据各维度间没有相关性,且网络传输较为正常时,算法具有明显优势.
For Skyline computation needs long time to deal with large amount of data,this paper use the P2P network, to allocate calcu- lation pressure to each node in the network. In the preprocessing, the proper data mapping method is adopted to increase the decision a- bility between data in the same node, and significantly reduce the amount of local computation. In the computation of global Skyline points,by peer-to-peer transmission in the network, new algorithm can minimize dataset of every node. Experimental results and theo- retical analysis shows,the new algorithm can down the amount of data in local node to 10%. When the dataset is very large,data＇s each dimension isn＇t relative with others＇, and net transmission is normal, the algorithm has obvious advantages.
出处
《小型微型计算机系统》
CSCD
北大核心
2017年第1期77-82,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61073037)资助
