High-Speed Private Information Retrieval Computation on GPU

A Private Information Retrieval (PIR) scheme is a protocol in which a user retrieves a record out of n from a replicated database, while hiding from the database which record has been retrieved, as long as the different replicas do not collude. A specially interesting sub-field of research, called single-database PIR, deals with the schemes that allow a user to retrieve privately an element of a non-replicated database. In these schemes, user privacy is related to the intractability of a mathematical problem, instead of being based on the assumption that different replicas exist and do not collude against their users. Single-database PIR schemes have generated an enormous amount of research in the privacy protection field during the last two decades. However, many scientists believe that these are theoretical tools unusable in almost any situation. It is true that these schemes usually require the database to use a lot of computational power, but considering the large number of applications these protocols have, it is important to develop practical approaches that provide acceptable performances for as many applications as possible. We present in this article a proof-of-concept implementation of a single-database PIR scheme proposed by Aguilar and Gaborit [2, 3]. This implementation can run in a CPU or in a GPU using CUDA, nVidia’s library for General Purpose computing on Graphics Processing Units (GPGPU). The performance results highlight that linear algebra PIR schemes allow to process database contents several orders of magnitude faster than previous protocols.