目录镜像及文件镜像算法

Catalog Mirroring and the File Mirroring Algorithm

本文介绍一种用软件使网络中的两台服务器实现某个目录镜像的方法。文件镜像算法是目录镜像的基础。算法找出源文件和目标文件中相同的部分 ,只传输不匹配的部分。算法不需要将两个文件放在一台机器上就能计算出它们之间的差别。当两个文件差别很小时算法效率最高 ,而差别较大时算法也保持正确和较高的效率。

This paper introduces a software method of mirroring some catalog on two networked servers. Catalog mirroring is based on the file mirroring algorithm, which identifies parts in the source file which are identical to those in the destination file, and sends only the parts which are matched. The algorithm computes a set of differences between the two files with both of them not necessarily on the same machine. It works the most effectively when the two files differ the least. However, it will still works in a rather correct and effective way even if the two files are quite different.