利用神经网络识别存储系统I/O特性  

I/O Characteristics Identification of Storage System by Neural Network

作　　者：王依凡 孙若琳 宗依林 刘春晓 孙志卓[1] WANG Yifan;SUN Ruolin;ZONG Yilin;LIU Chunxiao;SUN Zhizhuo(School of Computer and Information,Dezhou University,Dezhou,China,253023)

机构地区：[1]德州学院计算机与信息学院,山东德州253023

出　　处：《福建电脑》2024年第10期35-38,共4页Journal of Fujian Computer

基　　金：德州学院大学生创新创业训练计划项目(省级)《数据迁移节能优化——让大数据存储更加节能》(No.S202310448122)资助。

摘　　要：大数据推动了存储数据呈现爆炸性增长,导致存储系统的性能与能耗问题日益突出,因此需要对存储系统进行性能与节能优化。在存储系统优化中,存储系统的I/O特性识别是优化的基本依据。为此,本文提出一种基于神经网络的存储系统I/O特性识别方法,利用神经网络技术识别存储系统I/O请求的顺序性与随机性,以及存储系统的实时性能需求等特性。实验的结果表明,该方法能够高效准确识别存储系统的I/O特性。Big data has driven explosive growth in stored data,leading to increasingly prominent performance and energy consumption issues in storage systems.Therefore,it is necessary to optimize the performance and energy efficiency of storage systems.In storage system optimization,identifying the I/O characteristics of the storage system is the fundamental basis for optimization.To this end,this article proposes a neural network-based method for identifying the I/O characteristics of storage systems,utilizing neural network technology to identify the sequence and randomness of storage system I/O requests,as well as the real-time performance requirements of the storage system.The experimental results show that this method can efficiently and accurately identify the I/O characteristics of the storage system.

关 键 词：神经网络 I/O特性 存储系统优化 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]