Keyword Binning-Based Efficient Search on Encrypted Cloud Data

Cloud provides storage facility to its users at a reduced cost, but storing confidential data on the cloud is a security concern. To provide data privacy, the confidential data are encrypted before outsourcing to the cloud. Storing encrypted data in the cloud affects the utilization of data as it makes document searching inefficient. Although recent research has tried to make searching efficient, still there is a trade-off between the search efficiency and search accuracy. As the cloud is based on the pay-per-use model, more is the time required to retrieve the relevant documents more will be the financial burden on end users which affects their cloud-usage satisfaction. In this paper, the concept of keyword binning is proposed where document indexes are assigned to multiple buckets based on the contained keywords and search is performed only in relevant buckets. So, we use the number of comparisons as one of the metrics as it affects the search time. With ranking incorporated, the retrieved results are ranked efficiently and in 84% of the cases 4 out of the top-5 results of the proposed scheme match the 4 results in top-5 plain-text ranked results. For enhanced query privacy, we also propose an efficient query randomization scheme. The experimental results using Reuters-21578 dataset show that the proposed scheme is privacy-preserving and efficient with 100% recall and 98.45% precision.