Implementing a segmentation-based oblivious RAM
Wen Dong Zhao · Iowa State University Digital Repository (Iowa State University) · 2014
As cloud-based storage becomes more popular, an increasing amount of data has been outsourced to cloud storage. For many businesses, individuals, or even governments, data privacy has become an important concern. The confidentiality of data can be protected by encryption, but this alone may not suffice for ensuring data privacy. Data access patterns can leak a considerable amount of information about the data as well. Even though existing Oblivious RAM (ORAM) constructions are provable solutions to this issue, their performance in practice has been impeded by the high communication and storage overheads incurred. In this thesis, we implemented a Segmentation-based ORAM (S-ORAM) to reduce communication and storage overhead. Experiments have shown that S-ORAM can effectively protect user's privacy as well as achieve low communication and storage overhead.