Query processing over secure perturbed data over hybrid cloud

Abstract

PurposeHybrid cloud composing of public and private cloud is seen as a solution for storage of health care data characterized by many private and sensitive data. In many hybrid cloud-based solutions, the data are perturbed and kept in public cloud, and the perturbation credentials are kept in private cloud.Design/methodology/approachHybrid cloud is a model combing private and public cloud. Security for the data is enforced using this distribution in hybrid clouds. However, these mechanisms are not efficient for range query and retrieval of data from cloud. In this work, a secure and efficient retrieval solution combining K-mean clustering, geometric perturbation and R-Tree indexing is proposed for hybrid clouds.FindingsCompared to existing solution, the proposed indexing on perturbed data is able to achieve 33% reduced retrieval time. The security of indexes as measured using variance of differences was 66% more than existing solutions.Originality/valueThis study is an attempt for efficient retrieval of data with range queries using R-Tree indexing approach.