迁移实例状态监控及失败恢复机制  

Status monitoring and fault recovery scheme of migrating instances

作　　者：卢朝霞[1] 曾广周[2] 

机构地区：[1]山东体育学院计算机信息工程中心,济南250102 [2]山东大学计算机科学与技术学院,济南250061

出　　处：《计算机工程与应用》2010年第26期17-21,共5页Computer Engineering and Applications

基　　金：国家自然科学基金No.60573169~~

摘　　要：迁移工作流是一种多执行主体并发协同工作的工作流范型,其自组织、自协调特性决定了迁移实例的执行状态容易受到外部环境和内部执行状态的影响而发生异常。对异常状态的检测以及发生故障后的恢复是保障其可靠性的关键。然而目前已有的监控处理机制不能有效支持对此问题的处理,因此,提出一种基于层次监控模型的迁移实例失败的协调恢复机制,实现对多迁移实例执行状态的监控以及执行失败后实现状态的一致性恢复。初步的实验表明,该机制能有效监控具有并发协同行为的执行主体,并保障恢复的一致性。Migrating workflow is a workflow model with multi-executors working simultaneously and collaboratively, Migrating instances are susceptible to exterior environment and interior status since the characteristics of self-recognize and self-coordinate.The exception detector and fault recovery is Key to guarantee the reliability.However,existing monitoring and processing scheme cannot provide effective measures.To solve above problems, a fault recovery scheme of migrating instances based on hierarchical monitoring model is presented based on the background of migrating workflow model.The scheme can monitor the execution status of migrating instances and implement consistent recovery when a migrating instance fails.The elementary experiments show that the scheme not only can monitor executing agents with collaborating actions effectively,but also can recover to a consistent system status.

关 键 词：迁移工作流 迁移实例 监控 恢复 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]