海量存储系统可用性建模与分析  

Availability modeling and analyzing of massive storage system

作　　者：路辉[1] 温东新[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《计算机工程与设计》2017年第1期65-69,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61370085)

摘　　要：针对IBM XIV海量存储系统的可用性进行建模与分析。采用层次化建模方法,对存储系统上层模块采用可靠性框架进行建模,对底层模块使用连续时间马尔科夫模型进行建模;综合考虑故障率、维修率、故障检测时间、故障检测延迟、故障检测覆盖率、维修人员反应时间等多种因素对存储系统可用性的影响,提高可用性建模的精确度。通过模型计算工具求得存储系统稳态可用度和年故障停机时间,对影响存储系统的参数进行敏感性分析。An availability model of high availability massive storage system was presented. A hierarchical approach was used in which reliability block diagrams were used in upper levels and homogeneous continuous time Markov chains （CTMS） were used in the bottom level. Parameters such as failure rate, maintenance rate, fault detection time, fault detection coverage ratio and maintenance personnel response time were taken into account. The steady state availability and annual downtime were calculated through the proposed model solving tool and the sensitivity analysis for some parameters was carried out.

关 键 词：海量存储系统 可用性 模型计算 可靠性框图 马尔科夫模型 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]