摘要
针对视频网站数据优化存储的问题,定义了文件热度的概念,利用平均值的方法描述了各个类型数据的热度。在此基础之上采用时间序列方法对数据进行预测,得到了未来十天的不同类型视频的访问情况。最后,分析了文件多个周期的热度,通过隶属度函数确定出每一个周期的淘汰阈值,通过计算不同周期的淘汰阈值,得到一个动态的淘汰阈值示意图,从而给出文件淘汰分发的原则,实现了数据的存储管理优化。
This paper studies the data transmission speed of the video website.First,the heat of files is defined,and the heat degree of various types of data is also described with the average heat method.Then,the access of different videos over the next 10 days is received based on the time series prediction.At last,against the analyzed results of the file heat degree during multiple cycles,a phase-out threshold of each cycle is determined by the membership function.By calculating the eliminating threshold of different cycles,a dynamic eliminating threshold scheme is derived to provide the file distribution principle,consequently realize the management optimization of data storage.
出处
《沈阳工程学院学报(自然科学版)》
2014年第4期366-369,共4页
Journal of Shenyang Institute of Engineering:Natural Science
关键词
存储优化
平均分析法
时间序列预测
模糊数学
淘汰阈值
Storage Optimization
Average Analysis
Time Series Prediction
Fuzzy mathematics
Eliminate Threshold