云环境下两阶段安全驱动的容错调度算法  

Two-stage security-driven and fault tolerant scheduling algorithm in cloud computing environment

作　　者：白晶晶[1] 田生伟[2] 禹龙[3] 于炯[2] 田国忠[4] 

机构地区：[1]新疆大学信息科学与程工学院,新疆乌鲁木齐830046 [2]新疆大学软件学院,新疆乌鲁木齐830008 [3]新疆大学网络中心,新疆乌鲁木齐830046 [4]新疆工程学院计算机工程系,新疆乌鲁木齐830000

出　　处：《计算机工程与设计》2014年第9期3065-3069,3090,共6页Computer Engineering and Design

基　　金：新疆维吾尔自治区自然科学基金项目(2013211A010)

摘　　要：为解决云环境下安全调度和可靠性问题,综合考虑云计算共享性、动态性等特点,以具有依赖关系的并行任务为基础,提出一种两阶段安全驱动的容错调度算法(TSDFT)。建立安全模型计算任务调度风险率,根据风险率选择处理机,使用自适应备份策略对任务进行备份预处理;在上一阶段基础上,通过被动副本方式实现任务容错调度,每个处理机维护主/副本2个局部队列,支持一个以上处理机同时失效。仿真结果表明,该算法能有效降低异构系统中任务调度风险率,提高调度的安全性和可靠性。To solve the security and reliability problems in cloud computing environment, a two-stage security-driven and fault tolerant scheduling algorithm named TSDFT was proposed based on parallel tasks with precedence constraints, which took the common shared and dynamic characteristics of the cloud computing into account. Firstly, a security model was established to calculate the risk rate of the task scheduling. Then secure processors were selected in the light of the risk rate and the self-adaptive replication strategy was designed to copy tasks. Therefore the fault tolerant scheduling was finally achieved in the passive-back- up-copy way on the basis of the previous stage. Every processor maintained two local queues, namely the primary copy queue and the backup copy queue, which enabled TSDFT supported more than one proeessor＇s failure at the same time. Simulation results show that TSDFT is efficient in reducing the risk rate and improving the security and reliability of the task scheduling in heterogeneous systems.

关 键 词：云计算 任务调度 容错 安全性 可靠性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]