S3C: An architecture for space-efficient semantic search over encrypted data in the cloud
Jason Wolfgang Woodworth, Mohsen Amini Salehi, Vijay V. Raghavan · 2016
The recent rapid growth in Internet speeds and file storage requirements has made cloud storage an appealing option on both a personal and enterprise level. Despite the many benefits offered by cloud storage, many potential users with sensitive data refrain from fully utilizing this service due to valid concerns about information privacy. An established solution to this concern is to perform encryption on the user side with the key stored on a local machine, meaning the cloud will never see the user's plaintext data. However, by encrypting data on the user side data processing capabilities (e.g., searching) are lost. In particular, the ability to semantically search is of the user's interest in large datasets. In this paper, we present S3C, a system that provides a semantic search functionality over encrypted data in the cloud. S3C combines approaches from traditional keyword-based searchable encryption and semantic web searching. It offers a user transparent experience that accepts a simple multi-phrase query and returns a list of documents ranked by semantic relevance to the query. Our proposed approach is space-efficient, which makes it suitable for large scale datasets. Our minimal processing also allows the system to be run on thin clients such as smart-phones or tablets. We evaluate the performance of our system against various real-world datasets, and our results show that it produces accurate search results while maintaining minimal storage overhead (~0.3% of the dataset size).