Encrypted Dedup-Upload, Download and Update Securely With Proof of Ownership

Over the past decade, the Internet of Things (IoT) has experienced rapid expansion, and the storage of necessary files at the network edge has become increasingly prevalent through the use of data deduplication. Data deduplication increases network and storage efficiency by only retaining a single copy of duplicated data on the server. However, storing files without any encryption leads to data leakage. Hence existing models utilise Message-Locked Encryption (MLE). Nonetheless they are vulnerable to brute-force attacks. Meantime, these methods also use Proof of Ownership (PoW) approaches which mitigate duplicate faking attacks, but are susceptible to replay attacks or result in high computational costs. This work proposes an encrypted deduplication model that utilises enhanced PoW. There are 4 modules in the model: upload, download, subsequent upload and update. The model employs an inter edge server aided Randomised Convergent Encryption (RCE) method to encrypt files. Then, blocks are encrypted using an intra-edge edge-aided MLE technique. The PoW algorithm is implemented to prevent duplicate-faking and also one time token is used to prevent replay attacks. The model also incorporates verification of a cuckoo filter to increase security for PoW.