Leveraging Journaling File System for Prompt Secure Deletion on Interlaced Recording Drives

Shuo-Han Chen, Kuo-Hao Huang

IEEE Transactions on Emerging Topics in Computing(2023)

引用 1|浏览3
暂无评分
摘要
With the growing awareness of secure computation, more and more users want to make their digital footprints securely deleted and irrecoverable after updating or removing files on storage devices. To achieve the effect of secure deletion, overwritten-based secure deletion techniques have been proposed to overwrite invalidated storage space with scrambled data content. Nevertheless, overwritten-based secure deletion requires explicit write requests for rewriting invalidated storage space with random data, thus incurring additional internal write traffic on storage devices. Recently, this inefficient aspect of secure deletion is exacerbated by the emergence of high-density interlaced magnetic recording (IMR) technology. IMR technology enhances the storage density of hard disk drives by interlacing narrower top tracks on wider bottom tracks. Owing to the interlaced track layout, securely erasing invalidated data on bottom tracks interferes with adjacent top tracks and track rewrites are required to preserve valid data on top tracks; thus, overwritten-based secure deletion on IMR drives significantly enlarges the internal write traffic and impairs the secure deletion efficiency. In this paper, instead of directly alleviating the constraint induced by the interlaced track layout, we propose leveraging the interlaced track layout of IMR drives in combination with the journaling data stream of journaling file systems to enable prompt secure data deletion while minimizing the write traffic of secure deletion operations. Experimental results suggest that the performance improvement achieve by the proposed prompt secure deletion (P-SD) strategy is 51.84% on average, when compared with the previous TrackLace approach on IMR drives.
更多
查看译文
关键词
IMR,secure deletion,interlaced,HDD,journaling file system
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