Computing and Informatics

The data de-duplication system not only pursues the high de-duplication rate, which refers to the aggregate reduction in storage requirements gained from de-duplication, but also the de-duplication speed. To solve the problem of random parameter-setting brought by Content Defined Chunking (CDC), a self-adaptive data chunking algorithm is proposed. The algorithm improves the de-duplication rate by conducting pre-processing de-duplication to the samples of the classified files and then selecting the appropriate algorithm parameters. Meanwhile, FastCDC, a kind of content-based fast data chunking algorithm, is adopted to solve the problem of low de-duplication speed of CDC. By introducing de-duplication factor and acceleration factor, FastCDC can significantly boost de-duplication speed while not sacrificing the de-duplication rate through adjusting these two parameters. The experimental results demonstrate that our proposed method can improve the de-duplication rate by about 5 %, while FastCDC can obtain the increase of de-duplication speed by 50 % to 200 % only at the expense of less than 3 % de-duplication rate loss.