Analysis of Transmission and Storage Gains from Sliding Checksum Methods

摘要

In a previous report we described, modelled and analyzed a protocol 'rsync' for synchronizing related files at different ends of a communications channel with a minimum of transmitted data. This report identifies the extent of the gains that this method may provide by performing experiments on large repositories of data. The outcome is a new method of compressing large directories of data that together with conventional methods provides major gains in compression factors. As is explained in the report these gains in compression are also indicative of the data transmission gains available when mirroring a collection of files at a remote site using rsync.