摘要
针对资源池化的云数据中心资源共享和快速弹性的特点,进行了面向服务的任务可靠性建模。该模型不仅生动还原了资源池中功能单元和模块跨越所属物理机柜进行冗余的情况,而且利用全概率公式,引入逻辑关联函数,在多层前提条件下给出了系统严重故障平均间隔时间(MTBCF)的解析算法。进一步地,通过Markov状态转移图进行了动态建模,并在相应的运维策略假设下,分析了当资源池及输入端满足一定装备完好率时的稳态可用度。
Considering two traits-resources sharing and elastic computing of a cloud data center with resources pooled,service-oriented modeling on mission reliability is discussed.The model not only reflects vividly the fact that functional units and modules in the resource pool keeping trans-rack redundancy,but also according to total probability formula brings in a logical and dependent function,leading to an analytical algorithm for MTBCF in the system under multilevel conditions.Furthermore,the analysis of stable availability about the pool and the input side is made via Markov state transition model when a certain materiel readiness is given and a corresponding pattern of operation and maintenance is assumed.
作者
刘上
马辰
张小凡
Liu Shang;Ma Chen;Zhang Xiaofan(Department of Top-level Information and Electronics,China Ship Research and Development Academy,Beijing 100192,China)
出处
《质量与可靠性》
2020年第4期34-42,共9页
Quality and Reliability
