SLA感知的事务型组合服务容错方法  被引量:3

SLA-Aware Fault-Tolerant Approach for Transactional Composite Service

在线阅读下载全文

作  者:张俊娜[1] 王尚广[1] 孙其博[1] 杨放春[1] ZHANG Jun-Na;WANG Shang-Guang;SUN Qi-Bo;YANG Fang-Chun(State Key Laboratory of Networking and Switching Technology (Beijing University of Posts and Telecommunications),Beijing 100876, China)

机构地区:[1]网络与交换技术国家重点实验室(北京邮电大学),北京100876

出  处:《软件学报》2018年第12期3614-3634,共21页Journal of Software

基  金:国家自然科学基金(61472047;61571066)~~

摘  要:针对组合服务容错逻辑与执行逻辑不分离,以及容错过程易出现SLA(service level agreement)违反的现状,提出一种SLA感知的事务型组合服务容错方法.该方法首先采用有限状态机建模组合服务执行过程,对其状态进行监控;其次,采用监控自动机监控执行过程中的SLA属性,确保不出现SLA违反;然后,对于补偿过程,采用改进的差分进化算法快速寻找最优恢复规划;最后,该方法与组合服务执行逻辑相分离,所以易于开发、维护和更新.基于真实数据集的实验结果验证了所提方法在故障处理时间与组合最优度方面优于其他方法,并且对不同故障规模适应良好.Addressing the status quo that fault-tolerant logic of composite service is not separated from execution logic and service level agreement (SLA) violation appears frequently, this article proposes a SLA-based fault-tolerant approach for transactional composite services. Firstly, finite-state machine is adopted to model the execution process of the composite service and monitor the execution status. Secondly, monitoring automata is employed to monitor the SLA attributes during its execution to avoid SLA violation. Thirdly, an improved differential evolution algorithm is used to quickly determine the optimal recovery plan for the compensation process. Finally, a process is given to illustrate that as the approach is isolated from the execution logic of the composite service, it is easy to develop, maintain and update. The experimental results based on the real data sets show that the proposed approach is superior to other approaches in both the fault handling time and composition optimization. Meanwhile, the approach can deal with different fault scales.

关 键 词:组合服务 容错 服务级别协议(SLA) 差分进化算法 有限状态机 

分 类 号:TP311[自动化与计算机技术—计算机软件与理论]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象