Transparent File Deduplication with Reduced Update Cost on Encryption Enabled Mobile Devices.

Junbin Ren,Cheng Ji, Weiwei Jin, Chao Yu, Weichao Guo,Yajuan Du,Zongwei Zhu

International Conference on Parallel and Distributed Systems(2023)

引用 0|浏览3
暂无评分
摘要
Data deduplication has been long studied to achieve data reduction. However, deploying deduplication on encryption enabled mobile systems might consume much memory footprint and computation time for hash calculations. Moreover, frequent file updates on deduplicated files could badly degrade the deduplication efficacy due to the increased file-system metadata penalty. Considering the characteristics of mobile devices, an efficient data deduplication method is proposed in this paper. First, it separates the hash calculation into foreground and background stages. The background stage calculates the hash values of potentially duplicate files while the foreground stage quickly hashes the file which is being written using a lightweight hash algorithm. Second, a dual-level node structure is proposed to improve the file update efficacy for deduplicated files, saving more storage space against the file re-splitting. Besides, we implement a superlink call to make deduplication process compatible with file-based encryption. These methods are combined to realize a transparent file deduplication (TFDedup) approach, which eliminates redundant data and reduces the associated cost of file update. Experimental results show that TFDedup succeeds to lower the space consumption when serving file updates by 55.6% and accelerate the deduplication process by 50.3%.
更多
查看译文
关键词
Mobile Devices,Data Reduction,Mobile System,Hash Function,Storage Space,File System,Memory Footprint,Source Code,Time Cost,Instant Messaging,Data Sharing,Video Files,Data Block,Read Files,Time Overhead,Level Of Redundancy,Mb In Size,Flash Memory,Inner Nodes,Subsequent Updates,Computational Time Cost,Data Chunks,Smartphone
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要