Revisiting the Computational Practicality of Private Information Retrieval

The retrieval of information from a remote database server typically demands providing the server with some search terms to assist with the retrieval task. However, keeping the search terms private without undermining the server’s ability to retrieve the desired information is a common requirement for many privacy-preservingsystems. Private information retrieval (PIR) provides a cryptographicmeans for retrieving data from a database without the database or database administrator learning any information about which particular item was retrieved. In 2007, Sion and Carbunar considered the practicality of single-server computational PIR schemes and concluded that no existing construction is as efficient as the trivial PIR scheme— the server transferring its entire database to the client. While often cited as evidence that PIR is impractical, that paper did not examine multi-server information-theoretic PIR schemes, which are orders of magnitude more computationally efficient; further, a single-server lattice-based scheme by Aguilar-Melchor and Gaborit has recently been introduced, which is also much more computationally efficient than the schemes examined by Sion and Carbunar. In this paper, we report on a performance analysis of the above single-server lattice-based PIR scheme as well as two multi-server information-theoretic PIR schemes by Chor et al. and by Goldberg. Using analytical and experimental techniques, we find the end-to-end response times of these schemes to be one to three orders of magnitude (10–1000 times) smaller than the trivial scheme for realistic computation powers and network bandwidths. Our result indicates that private information retrieval can indeed be a practical tool for the protection of query privacy in database systems.