一种高效的合作实时事务并行检验点算法  

An Efficient Concurrent Checkpointing Algorithm for Distributed Cooperative Real-Time Transaction Processing

作　　者：李国徽[1] 王洪亚[1] 刘云生[1] 

机构地区：[1]华中科技大学计算机学院,武汉430074

出　　处：《计算机科学》2005年第7期69-71,75,共4页Computer Science

基　　金：青年国家自然科学基金(编号:60203017)

摘　　要：许多数据和活动上都有很强时间性的应用在地理上同时具有分布性,这种应用需求使得分布式实时数据库的研究成为数据库研究领域的热点。在实时事务执行时,事务故障或数据竞争会导致事务重启,为了减少因重启而损失的工作量,可以采用检验点技术以利于事务时间正确性的满足。在一些分布式实时数据库应用中,不同结点的事务通过消息交换形成合作关系,当某一事务记检验点时,为保证合作事务间的全局一致性,相关事务也要相应地记检验点。传统的协同检验点方法没有考虑应用的定时约束,不能很好地支持分布式实时事务处理。本文提出了一种高效的并行协同检验点方法,该算法既具有最小协同检验点特性又使全局检验点过程延时最小。实验表明该算法减少了全局检验点阻塞时间,有利于分布式实时事务截止期的满足。Distributed real-time database systems have gained more and more research interests in the last decade. When transactions have to restart due to transaction failures and data contention, taking checkpoints periodically can reduce the amount of lost work and help real-time transactions to meet their deadlines. In some distributed real-time database applications, transactions in different sites can communicate with each other by message exchange. To main- tain the consistency among transactions, when a transaction initiates a checkpoint, the elated transactions are forced to take their own corresponding checkpoints. Traditional coordinated checkpoint methods don't take the timing con- straints of applications into consideration and are unsuitable for supporting distributed real-time transaction processin- g. This paper proposes a concurrent coordinated checkpointing method, which minimizes the time latency involved in a global checkpoint and force a minimal number of transactions to take their checkpoints. Extensive experiments show that the proposed algorithm results in the reduced time latency and an obvious reduction in the transaction deadline- missing rate as well.

关 键 词：检验点 算法 高效 分布式实时数据库 分布式实时事务 实时事务处理 数据库应用 研究领域 应用需求 数据竞争 合作关系 消息交换 合作事务 并行协同 分布性 时间性 多数据 工作量 正确性 一致性 点过程 截止期 最小 结点 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP301.6[自动化与计算机技术—计算机科学与技术]