On Insecure Uses of BGN for Privacy Preserving Data Aggregation Protocols

TL;DR

This paper critically examines the privacy vulnerabilities of BGN cryptosystem-based data aggregation protocols, revealing that existing blinding techniques can leak information to decryptors due to the pairing-based structure, and proposes mitigation strategies.

Contribution

It identifies privacy flaws in the BGN cryptosystem's blinding technique used in data aggregation and offers solutions to prevent such leakage.

Findings

BGN-based protocols are vulnerable to privacy leakage from decryptors.

The pairing e:GxG-->G_T makes DDH problem on G easy, compromising privacy.

Proposed methods to enhance privacy in BGN cryptosystem applications.

Abstract

The notion of aggregator oblivious (AO) security for privacy preserving data aggregation was formalized with a specific construction of AO-secure blinding technique over a cyclic group by Shi et al. Some of proposals of data aggregation protocols use the blinding technique of Shi et al. for BGN cryptosystem, an additive homomorphic encryption. Previously, there have been some security analysis on some of BGN based data aggregation protocols in the context of integrity or authenticity of data. Even with such security analysis, the BGN cryptosystem has been a popular building block of privacy preserving data aggregation protocol. In this paper, we study the privacy issues in the blinding technique of Shi et al. used for BGN cryptosystem. We show that the blinding techniques for the BGN cryptosystem used in several protocols are not privacy preserving against the recipient, the decryptor. Our analysis is based on the fact that the BGN cryptosystem uses a pairing e:GxG-->G_T and the existence of the pairing makes the DDH problem on G easy to solve. We also suggest how to prevent such privacy leakage in the blinding technique of Shi et al. used for BGN cryptosystem.