A Secure and Dynamic Multi-keyword Ranked Search Scheme over Encrypted Cloud Data

Abstract

Due to the expanding prevalence of distributed computing, an ever increasing number of information proprietors are spurred to outsource their information to cloud servers for extraordinary comfort and decreased cost in information administration. Notwithstanding, delicate information ought to be encoded before outsourcing for security necessities, which obsoletes information usage like catchphrase based report recovery. In this paper, we exhibit a protected multi-watchword positioned seek plot over scrambled cloud information, which all the while underpins dynamic refresh tasks like cancellation and addition of reports. In particular, the vector space demonstrates and the broadly utilized TF_IDF display is consolidated in the file development and question age. We build an uncommon tree-based record structure and propose an "Insatiable Depth-first Search" calculation to give proficient multi-catchphrase positioned look. The protected KNN calculation is used to encode the record and inquiry vectors, and in the meantime guarantee exact importance score figuring between scrambled list and question vectors. With a specific end goal to oppose measurable assaults, ghost terms are added to the file vector for blinding list items. Because of the utilization of our unique tree-based list structure, the proposed plan can accomplish sub-straight hunt time and manage the erasure and inclusion of reports adaptably. Broad analyses are directed to exhibit the productivity of the proposed plot.