DLSBD-MHT: Dual-level source-based deduplication with Merkle-Hash-Tree for big data
Ke Huang, Xiaosong Zhang, Lin Jun Sun, Xiaofen Wang · 2017
Big data grows fast which causes significant strain to storage. Deduplication is a practical technique to decrease storage load by deleting repeated data. However, it faces with efficiency and security issues. Efficiency is the first to consider since deduplication becomes pointless if it poses burden to cloud storage. Meanwhile, current deduplication schemes suffer from both inside and outside attacks. Concretely, malicious users may cheat to obtain ownership of data file or cause damage to it, and the server may extract privacy information from data for political or economic use. To solve this problem, we propose a Dual-Level Source-Based Deduplication with Merkle-Hash-Tree for big data, called DLSBD-MHT. Our scheme is built on Merkle-Hash-Tree (MHT) and Block-Level Message-Locked Encryption (BL-MLE). It can securely and reliably avoid identical upload, thus it increases the efficiency of big data in a safe way. We provide security proof and simulation to validate our proposal. The evidence shows our scheme is secure and efficient in use.