Fast Information Retrieval over Encrypted Outsourced Cloud Data

Abstract

The data used in cloud applications is directly exposed to the cloud service provider, and because of the potential compromise of the cloud, could also be learned by adversaries. When encrypted data is hosted on cloud provided that there are large amount of data files, utilization of encrypted data effectively becomes a very challenging task. In a cloud computing environment, where outsourced data of organizations is shared with a large number of users. These variety of users might be interested in retrieving certain specific data files during a given session. A popular and interesting way to do so is by using keyword-based search. These search techniques facilitate users to search and retrieve data files selectively in which the users are interested. These keyword-based searches are being widely used for plaintext searches. But data encryption poses a challenge to perform keyword search using existing paintext search methods to be used for encrypted outsourced data on cloud. In this paper, we have analyzed the searchable indexes that could be used to make a fast and effective search on encrypted outsourced data and proposed a scheme that could make fast and accurate searches over encrypted outsourced cloud data. Simulation results have revealed that the proposed scheme takes much less time in generating the searchable index as compared to already existing techniques. The vector space model being used earlier for keyword based searches on encrypted data, is relatively time consuming and hence leads to very high time complexity during relevance score calculations as well as index generation for large datasets. Hence the proposed scheme achieves a fast and secure relevance scoring for large number of datasets also and in much less time as compared to the vector space model.