A NOVEL BLOOM SEARCH INFORMATION RETRIEVAL (BSIR) MODEL FOR MULTI-KEYWORD SEARCH OVER THE CLOUD WITH HIGH SECURITY & DATA FRESHNESS

Abstract

The growing acceptance of cloud computing has provided strong incentives for data owners to transfer private information to remote servers. Privacy is a significant issue when outsourcing data to the cloud. Data owners commonly employ encryption techniques to protect the privacy of their documents before outsourcing them. Given the exponential growth of data, it is imperative to establish effective and dependable methods for cipher text search. This will enable data owners to retrieve and modify cloud data conveniently. This research presents a novel framework named the Bloom Search Information Retrieval (BSIR) model. This model aims to enable a secure multi-keyword ranked search scheme while also ensuring data integrity. Data integrity is maintained through measures that address the data's completeness, correctness, and freshness. A verification set (AS) is also created to validate the top-k results. The proposed BSIR model combines the vector space and TFIDF framework, commonly used for index creation and query generation. Comparison with other models (BDMRS, EDMRS, KNN) demonstrates that the BSIR model is highly secure and efficient. Our scheme optimizes search efficiency and minimizes communication overhead in verifying search results.  The time for verifying search results for 10 documents for Proposed BSIR model is 3.4 seconds, and time for searching files in the cloud with 100 documents is 48 seconds and the time for index construction for 50 documents is 22 seconds, time for generating trapdoor is 5 seconds for 25 keywords.