摘要
对计算系统执行智能化细粒度软件抗衰策略可以进一步增强软件抗衰技术的适用性,更大程度地提高抗衰效率,进一步降低抗衰成本,提高软件可靠性。通过分析系统进程的相关数据,针对进程间控制、调用及数据访问的关系,定义了进程间的依赖程度,给出了进程间重启相关度和进程重启链的判定步骤以及进程最优恢复策略制定准则;并在此理论基础上结合Agent工作原理及特点,构建了基于agent的软件抗衰体系结构,并以此判定了系统中任意进程间的重启相关度,生成了进程重启链,并制定了重启进程的最优抗衰策略,由此从根本上实现了计算系统的智能化细粒度软件抗衰。
In order to improve the availability and efficiency of software rejuvenation, decrease the rejuvenation cost and enhance the reliability of software, the intelligent software rejuvenation with fine granularity was executed. Based on analyzing the relative data of processes and the coupling relation between processes of software systems, the degree of coupling relation was defined between the processes, qnd determination steps of the degree of restart dependence and restart chains and establishing rule of the best process software rejuvenation policy were put forward. Integrating with the principium and characters of agent, a software system was set down based on multi-agents which determined the degree of restart dependence ,restart chain and the best software rejuvenation policy of process, so that the intelligent software rejuvenation with fine rejuvenation granularity was achieved radically.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2007年第23期5392-5396,共5页
Journal of System Simulation
基金
国家自然科学基金项目名称(60273035)
国防科工委基础应用项目(K1704060511)
关键词
软件抗衰
重启相关度
重启链
AGENT
software rejuvenation
degree of restart dependence
restart chain
agent