摘要
为了对抗软件在运行过程中出现的老化现象,需要实施合适的软件恢复(software rejuvenation)策略。该文将单纯基于时间的和基于检测的软件恢复策略相结合,提出了基于时间与检测的软件恢复策略,弥补了前者确定的恢复时间间隔较为保守和后者引发过高的监控成本的缺陷。对3种策略的成本分析和比较表明,文章在引发较小监控成本的前提下,进一步延长了恢复时间间隔,从而降低了恢复策略的总的实施成本。
To counteract software aging during the running time of system, it is essential to bring a proper software rejuvenation policy into effect. The conventional time-based software rejuvenation policy （Policy Ⅰ ）can be implemented simply with less interval, which incurs higher rejuvenation cost. While the detection-based rejuvenation policy （Policy Ⅱ ）can rejuvenate timely but incur higher detection cost. Therefore a software rejuvenation policy based on time ＆ detection （Policy Ⅲ） is put forward in this paper for overcoming these defects. The numerical results of analyzing and comparing the mean time cost of three policies demonstrate that the policy m prolongs rejuvenation interval without too much detection cost, consequently decreases total implementation cost.
出处
《计算机工程》
CAS
CSCD
北大核心
2006年第11期42-43,49,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60273035)
国防科工委基础应用项目
关键词
软件恢复
系统可靠性
软件老化
Software rejuvenation
System reliability
Software aging
