Aiding secure data retrieval incorporated with parallelization technique in cloud

Abstract

<p>Cloud data owners prefer to outsource data because of ease in maintenance. Data confidentiality of this outsourced sensitive data is a major task. The searchable encryption technique helps to carry out searches on encrypted data without decrypting it. In the data outsourcing environment, volume of data is increasing rapidly. Hence the time required to build the index and to carry out searches is also increasing exponentially. This makes it more difficult to build a system which is efficient, reliable and can cope up with growing data. In this paper, a parallelization technique to build the index on outsourced data is proposed. This technique minimizes the time required to construct the index. It also supports secure ranked retrieval using bucketization technique. The buckets are formed using Hadoop map reduce framework which achieves significant efficiency. The proposed method prune the keyword dataset, which helps in significant reduction in the size of index. Through extensive experiments using standard dataset, the performance of the system is validated. The experimental results show that the proposed system requires less time for index construction and hence improves retrieval efficiency.</p>