基于预测模型和缓存替换策略的网络资源访问研究

发布时间：2018-05-05 02:37

本文选题：制造资源 + 预取技术　；参考：《科技通报》2017年10期

【摘要】：针对在网络制造资源共享中客户端对资源无法有效、快速访问的问题进行应用研究,提出基于预测的网络制造资源访问策略,解决网络延迟问题。首先,描述了在客户端如何进行局部Cache设计,然后描述了在传统替换算法的基础上,如何引入基于多项式回归的预取技术,如何将预取技术与语义缓存技术相结合。通过构建对用户查询频率进行实时监测和基于多项式回归预取技术的预测模型,实现动态在线预测,避免了用户兴趣漂移引起的预取不确定性,较好地解决了访问延迟问题。最后,设计相关实验对客户端访问资源的有效性、快速性进行验证。实验结果表明,该优化调度策略在命中率上相对于传统的替换算法大约提高了8%,在一定的缓存空间内较传统替换算法在性能上有一定的提高,提高了客户端访问资源的有效性、快速性。
[Abstract]:In order to solve the problem that the client can not access the resource effectively and quickly in the network manufacturing resource sharing, a predictive based access strategy is proposed to solve the problem of network delay. Firstly, this paper describes how to design the local Cache in the client, then describes how to introduce the polynomial regression prefetching technology based on the traditional replacement algorithm, and how to combine the prefetching technology with the semantic cache technology. Based on the real-time monitoring of user query frequency and the prediction model based on polynomial regression prefetching technology, dynamic online prediction is realized, which avoids the uncertainty of prefetching caused by user interest drift, and solves the problem of access delay. Finally, some experiments are designed to verify the efficiency and rapidity of client access. The experimental results show that the hit rate of the optimal scheduling strategy is about 8% higher than that of the traditional replacement algorithm, and the performance of the optimized scheduling strategy is improved compared with the traditional replacement algorithm in a certain cache space, and the efficiency of accessing the resources on the client side is improved. Rapidity.
【作者单位】：四川机电职业技术学院;
【分类号】：TP333

【相似文献】