Privacy-aware relevant data access with semantically enriched search queries for untrusted cloud storage services

Zeeshan Pervez, Mahmood Ahmad, Asad Masood Khattak, Sungyoung Lee, Tae Choong Chung

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)
136 Downloads (Pure)


Privacy-aware search of outsourced data ensures relevant data access in the untrusted domain of a public cloud service provider. Subscriber of a public cloud storage service can determine the presence or absence of a particular keyword by submitting search query in the form of a trapdoor. However, these trapdoor-based search queries are limited in functionality and cannot be used to identify secure outsourced data which contains semantically equivalent information. In addition, trapdoor-based methodologies are confined to pre-defined trapdoors and prevent subscribers from searching outsourced data with arbitrarily defined search criteria. To solve the problem of relevant data access, we have proposed an index-based privacy-aware search methodology that ensures semantic retrieval of data from an untrusted domain. This method ensures oblivious execution of a search query and leverages authorized subscribers to model conjunctive search queries without relying on predefined trapdoors. A security analysis of our proposed methodology shows that, in a conspired attack, unauthorized subscribers and untrusted cloud service providers cannot deduce any information that can lead to the potential loss of data privacy. A computational time analysis on commodity hardware demonstrates that our proposed methodology requires moderate computational resources to model a privacy-aware search query and for its oblivious evaluation on a cloud service provider.
Original languageEnglish
Article numbere0161440
JournalPLoS ONE
Issue number8
Publication statusPublished - 29 Aug 2016


Dive into the research topics of 'Privacy-aware relevant data access with semantically enriched search queries for untrusted cloud storage services'. Together they form a unique fingerprint.

Cite this