一种分阶处理的软件自恢复方法  

A Staging Processing Software Self-recovery Method

作　　者：杨光 董玺坤 王丹[2] YANG Guang, DONG Xi-kun, WANG Dan （1.716 Research Institute of China Shipbuilding Industry Coproration, Lianyungang 222006, China; 2. Harbin Engineering University, College of Computer Science and Technology, Harbin 150001）

机构地区：[1]中船重工第七一六研究所,江苏连云港222006 [2]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《电脑知识与技术》2010年第7期5245-5248,共4页Computer Knowledge and Technology

基　　金：国家自然科学基金重大计划（90718003）;国家自然科学基金（60373000）; 国家高技术研究发展计划（863）（2007AA01Z401）

摘　　要：任务关键系统要求高可用性,因此当系统发生故障时需要快速恢复。由于任务关键系统中的故障多与软件相关,为满足恢复时间短、对系统副作用小等任务关键系统的恢复需求,提出一种分阶处理的软件自恢复方法。在预处理过程中,采用所提出的微重启技术自动修改二进制文件使其在出现异常时重启恢复成为一种自律行为;在系统运行时,采用所提出的基于MD5监测算法及热插拔技术的软件自修复方法对系统实施监测及自动修复。该方法可恢复系统内部失效如响应超时、资源泄漏等问题;对系统受到外部攻击的情况,如病毒修改二进制文件,非授权用户非法篡改系统文件等,该方法也可有效动态恢复,同时可以兼容系统动态升级需求。Mission-critical systems require rapid recovery when failure occurs with consideration of high availability. To meet the requirement of short recovery time and fewer side effects on the mission-critical systems, this paper puts forward a software self-recovery method with staging process since much failure of mission-critical systems is related to the software. In the preprocessing procedure, the proposed microreboot technique is utilized to modify the binary files automatically to make their recovery be an autonomic action when anomaly occurs. During the running-time, the proposed software self-healing technique based on MD5 monitoring algorithm and hot-swapping is adopted to monitor and heal the system. The method can recover the inner failure of the system such as overtime response, resources leak, etc. And for the situation of being attacked from outside the system, such as malicious modification by viruses, illegal tamper by the unauthorized users, the method can also implement effective recovery. Meanwhile, it offers the dynamical upgrade ability for the system.

关 键 词：任务关键系统 自恢复 微重启 热插拔 分阶处理 

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