摘要
间歇故障具有随机性,复杂的机理与表现会导致传统诊断方法不再有效,尤其是对于云计算环境,因为长时间运转、高负载运行以及集群规模大等因素致使间歇故障易感性居高.针对此问题,提出了一种基于跨层级间歇故障传播行为的建模与量化分析方法.该方法对间歇故障行为演化进行跟踪并量化分析,包括发现并记录故障在云环境不同层级的传播,构建传播路径,量化间歇故障对结构级、固件、超特权、特权和用户层级的影响,以及故障模型参数对系统影响的比较,以为云环境下间歇故障诊断策略提供辅助诊断信息和指导.
Abstract
Intermittent faults have randomness,and complex mechanisms and manifestations can make traditional diagnostic methods no longer effective,especially for cloud computing environments,as factors such as long-term operation,high load operation,and large cluster size make intermittent faults highly susceptible.A modeling and quantitative analysis method based on cross-layer intermittent fault propagation behavior is proposed to address this issue.This method tracks and quantitatively analyzes the evolution of intermittent fault behavior,including discovering and recording the propagation of faults at different levels in the cloud environment,constructing propagation paths,quantifying the impact of intermittent faults on structure level,firmware,super privilege,privilege,and user level,as well as comparing the impact of fault model parameters on the system,providing auxiliary diagnostic information and guidance for intermittent fault diagnosis strategies in the cloud environment.
NETL
NSTL
万方数据