Open AccessEfficient multi-keyword similarity search over encrypted cloud documents
Author(s) -
Ayad I. Abdulsada,
Dhafer G. Honi,
Salah Al-Darraji
Publication year - 2021
Publication title -
indonesian journal of electrical engineering and computer science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.241
H-Index - 17
eISSN - 2502-4760
pISSN - 2502-4752
DOI - 10.11591/ijeecs.v23.i1.pp510-518
Subject(s) - computer science , encryption , outsourcing , similarity (geometry) , cloud computing , scheme (mathematics) , hash function , symmetric key algorithm , search engine indexing , information retrieval , data mining , nearest neighbor search , database , computer security , public key cryptography , artificial intelligence , mathematics , image (mathematics) , mathematical analysis , political science , law , operating system
Many organizations and individuals are attracted to outsource their data into remote cloud service providers. To ensure privacy, sensitive data should be encrypted be-fore being hosted. However, encryption disables the direct application of the essential data management operations like searching and indexing. Searchable encryption is acryptographic tool that gives users the ability to search the encrypted data while being encrypted. However, the existing schemes either serve a single exact search that loss the ability to handle the misspelled keywords or multi-keyword search that generate very long trapdoors. In this paper, we address the problem of designing a practical multi-keyword similarity scheme that provides short trapdoors and returns the correct results according to their similarity scores. To do so, each document is translated intoa compressed trapdoor. Trapdoors are generated using key based hash functions to en-sure their privacy. Only authorized users can issue valid trapdoors. Similarity scores of two textual documents are evaluated by computing the Hamming distance between their corresponding trapdoors. A robust security definition is provided together withits proof. Our experimental results illustrate that the proposed scheme improves thesearch efficiency compared to the existing schemes. Further more, it shows a high level of performance.