摘要
针对终端设备产生的数据上传至云服务器会产生时间延迟的问题,提出了基于UCBM算法的缓存预取优化策略。根据用户的访问特征,通过贝叶斯网络和马尔可夫链对用户的下一步访问行为进行分类与预测,将目标文件提前从云端缓存预取到最近的边缘服务器,从而减少了时间延迟。实验结果表明,提出的优化策略明显提高了文件的预取精确率,最高为0.78;覆盖率也明显优于已有算法,最高达到0.73。
To solve the problem of a long time delay for transferring data from terminal devices to the cloud server,this paper proposes a cache prefetching optimization strategy based on UCBM algorithm.According to user access characteristics,the next access behavior of users is classified and predicted by Bayesian network and Markov chain.Then target files are prefetched from the cloud cache to the nearest edge server ahead of time,reducing time delay.Experimental results show that the cache prefetching algorithm proposed in this paper improves the accuracy of file prefetched obviously,with a maximum prefetching accuracy rate of 0.78,and the coverage is also significantly superior to the exsiting algorithm,with a maximum coverage rate of 0.73.
作者
孙波
郭祖华
SUN Bo;GUO Zuhua(School of Computer Science and Technology,Henan Institute of Technology,Xinxiang 453003,China;Teaching quality monitoring and Evaluation Center,Henan Institute of Technology,Xinxiang 453003,China)
出处
《河南工学院学报》
CAS
2021年第5期12-18,共7页
Journal of Henan Institute of Technology
基金
河南工学院教育教学改革研究与实践重点项目(2019JG-ZD006)。
关键词
边缘计算
时延
贝叶斯网络
马尔可夫链
缓存预取
edge computing
time delay
Bayesian network
Markov chain
cache prefetching
