Research And Design Of Management-oriented Opening Universal Archive Tool

With the increasing user demands of data backup, the functions of file archive tool contains no longer just file copy, but the management of the archive file and other different data transforming. Nowadays, many mature backup systems are powerful, such as WinRAR and WinZip. However, most of them have some common drawback: The design of backup system is complicated, lacking of a universal backup model; and it also lacks of the management of the archive file set. It's difficult to maintain the backup system. Owing to its non support to the third-party extension, the development of backup system is high cost.The pattern of file archive tool is stream processing, the size of data to be processed is huge, and the function of each data processing is different with others. Supposing each data-processing is an independent entity and to filter and transform the data block, all of the data-processing can be coupled loosely, and the versatility and scalability of the system are enhanced. So in this article, we try to build a model which is based on the famous pipe-filter model. In this model, each data processing follows the same interfaces as a separate entity, which is linked up by the pipes to implement the file archive system.Base on some specific problems in popular backup systems, first we give some definition of the archive system in a semi-formal method, analyze and summarize its mathematical nature, build a streaming driven computational model. Under this model, we design a file archive system and prove the feasibility and correctness of using it .After that, we implement the system in pure C++ language. Finally, we compare our file archive system with other popular file backup system in both functions and efficiency, and summarize its deficiency.