A dynamic and verifiable multi-keyword ranked search scheme in the P2P networking environment

• Published in: Peer-to-peer networking and applications Vol. 13; no. 6; pp. 2342 - 2355
• Main Authors: Wang, Haoyang, Fan, Kai, Li, Hui, Yang, Yintang
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.11.2020; Springer Nature B.V
• Subjects: Algorithms; Cloud computing; Communications Engineering; Computer Communication Networks; Data storage; Encryption; Engineering; Information Systems and Communication Service; Modular structures; Networks; Peer to peer computing; Performance evaluation; Searching; Security; Signal,Image and Speech Processing; Social networks; Special Issue on Security and Privacy in Machine Learning Assisted P2P Networks; Merkle tree; Forward and backward security; Time-stamp chain; P2P network; Tree-based index; Secure kNN algorithm
• ISSN: 1936-6442; 1936-6450
• DOI: 10.1007/s12083-020-00912-7

With the rapid development of the Internet, Peer to Peer(P2P) network has been applied in various fields. Users in P2P network also have a large amount of data, but users cannot provide enough storage space locally. More and more users choose to upload their own data to cloud server in order to save overhead and facilitate sharing their own data with other users. In order to ensure data security, researchers have proposed searchable encryption(SE) technology, and searchable encryption has been widely used. In this paper, a dynamic verifiable multi-keyword ranked search scheme is proposed under the background of P2P network and cloud storage service(CSS). On the basis of using secure kNN algorithm to encrypt index and traditional inner product algorithm to obtain ranked results, the scheme in this paper realizes forward and backward security by changing the structure of file vector and using modular residual computation. Meanwhile, the integrity and freshness of search results are verified by combining timestamp chain and Merkle tree. Finally, the security of this scheme under two threat models is analyzed, and the performance evaluation experiment is carried out on the document set.