摘要
提出一种在网络备份系统(network backup system,NBS)中的海量缓存方法。给出海量缓存的定义、使用目的和与备份代价密切相关的大小评估,并描述了一种基于MS Windows海量缓存设计与实现的方法,并对其性能进行了实验分析。由于网络传输与备份中心处理的延迟,使得本地I/O系统处理速度远远大于数据备份的速度,从而造成本地内存缓冲队列的大小越来越大。海量缓存方法就是通过借助大量的辅存来解决这个问题,当本地数据中心监控到数据变化后,会先把数据变化放入辅存中,然后再发送到远程备份中心。
This paper designed mass buffer in network backup system (NBS), provided the definition, purpose, size assessment related to the backup cost, the design and implementation of mass buffer. Because of the delay caused by network transmission and process in remote backup center, the I/O rate in the local host is much faster than that of data backup, which caused the memory buffer queue will be more and more large. This problem was solved by the method of mass buffer, which was supported by secondary storage. The changes of data would be first put into mass buffer, from which the data then would be send to remote backup center.
出处
《计算机应用研究》
CSCD
北大核心
2009年第1期9-10,29,共3页
Application Research of Computers
基金
国家"863"计划资助项目(2006AA01Z435)
国家自然科学基金资助项目(60573130
60502011)
国家教育部博士点基金资助项目(20070610032)
关键词
海量缓存
容灾
内核
mass buffer
disaster tolerant
kernel