網路中之高效能檔案備份架構

Abstract

隨著網路技術及設備逐日普及，越來越多的資訊藉由網路所傳遞，亦有許多的『雲端服務』快速崛起。為了在提供服務時，能確保使用者資料的完整性，我們將於一個遠端系統上儲存資料的備份；而要提高此備份資料的可靠性及可還原性，則可藉由改良此同步過程的效率來實現。

在此論文中，我們將單一檔案切割為可變大小的複數資料區塊，在檔案變動時，我們藉由字串比對的方式判斷各區塊是否變動，並在比對結束後只同步變動後的區塊；如此便可減少所需同步的資料量，達到降低傳輸損耗的成本並提高整體備份過程的效率。

Because of the advance of technology and the popularization of network device construction, more and more information delivered by the Internet. And so the number of Cloud Services increased rapidly. To ensure the integrity of user data while providing services, the system should store backup data on remote storage systems. And we can improve the data reliability and recovery ability by improving the efficiency of synchronization processing.

In this paper, we propose a solution based on dividing a file into flexible size blocks. When the file modified, we can judge each block is changed or not by utilizing pattern matching algorithm, and only need to update those changed blocks into backup storage systems after matching. So we can lower the transmission cost and increase the efficiency of the whole backup processing by reducing the amount of data we need to transmit.