A Secondary Index for Improving Reading Performance in the Inline Deduplication System
Shuai Long, Yaping Lin, Xin Yao, Wei Zhang · 2016
With the advent of cloud computing, huge amount of data are stored on the cloud. To remove the redundant data, the Deduplication technology is proposed and attracts many interests from both the academic and industry. In the Inline Deduplication System, improving the read performance is very vital. However, existing works is inefficient when massive data are involved. In this paper, we first propose a Secondary Index Assisted Read scheme (SIAR) in the Inline deduplication system. To reduce the frequency of disk access, and improve the read performance, we build a secondary index in the RAM and disks, which makes full use of the high random-read performance properties of RAM and the low cost of disk. Further, we analyze the trade-off between the performance improvement of SIAR and memory overhead, which makes the proposed scheme adaptive to different applications. Finally, we conduct extensive experiments, which confirms the efficiency and efficacy of SIAR.