h(k)‐private information retrieval from privacy‐uncooperative queryable databases

Abstract

Purpose-This paper aims to address the privacy problem associated with the use of internet search engines. The purpose of the paper is to propose and validate a set of methods and protocols to guarantee the privacy of users' queries. Design/methodology/approach-In this paper h(k)-private information retrieval (h(k)-PIR) is defined as a practical compromise between computational efficiency and privacy. Also presented are h(k)-PIR protocols that can be used to query any database, which does not even need to know that the user is trying to preserve his or her privacy. Findings-The proposed methods are able to properly protect the privacy of users' queries. When internet users apply the protocols, search engines (e.g. Google) are not able to determine unequivocally the real interests of their users. The quality of the results does decrease with the increase in privacy, but the obtained trade-off is excellent. Practical implications-Current private information retrieval (PIR) protocols suffer from two significant shortcomings: their computational complexity is O(n) where n is the number of records in the database, which precludes their use for very large databases and web search engines; and they assume that the database server cooperates in the PIR protocol, which prevents deployment in real-life uncooperative settings. The proposed protocols overcome both problems. Originality/value-This is the first set of protocols that offer practical protection for the privacy of the queries that internet users submit to an internet search engine. The proposal has been implemented and it will be released to the general public soon. It will help to protect the right to privacy of millions of internet users.

SimAttack: private web search under fire

“…For instance, GooPIR [6] adds to the initial query (k − 1) fake queries (generated using a dictionary) where all of these queries are separated by the logical OR operation in a new obfuscated query. As GooPIR's authors consider that an adversary has no background knowledge about the user, this adversary can only guess the initial query with a probability equal to 1/k.…”

“…For instance, GooPIR [6] adds extra queries to the original query while TrackMeNot [7] sends periodically fake queries.…”

Coprivacy: Towards a Theory of Sustainable Privacy

“…Pragmatic approaches to guarantee some query privacy have therefore been based so far on two relaxations of PIR: standalone and peer-to-peer (P2P). In the standalone approach, a program running locally in the user's computer either keeps submitting fake queries to cover the user's real queries (TrackMeNot, [12]) or masks the real query keywords with additional fake keywords (GooPIR, [7]). In the P2P approach, a user gets her queries submitted by other users in the P2P community; in this way, the database still learns which item is being retrieved, but it cannot obtain the real query histories of users, which become diffused among the peer users, thereby achieving user-private information retrieval (UPIR).…”

Private Web Search with Constant Round Efficiency

“…Different approaches (e.g., [10,21,8,14,20,13]) have been proposed to tackle this problem. In these systems, the main measure of efficiency is the round complexity, and it is important to construct constant-round PWS systems while guaranteeing privacy.…”

“…Another approach to provide privacy during web search is based on a query obfuscation technique (e.g., [10,8,18,3]). Roughly speaking, a class of solutions using query obfuscation is to blend the real queries into a stream of fake queries so that web search engines cannot create a correct profile.…”