Efficient Weight-based Private Matching for proximity-based mobile social networks

Abstract

Proximity-based mobile social networks (PMSNs) are becoming increasingly popular in recent years with the explosive growth of mobile devices, where a user can find a best matching friend in vicinity through profile matching. However, the matching process calls for the publication of users' personal information, which conflicts with users' growing privacy concerns about revealing their profiles to strangers. To achieve privacypreserving friend discovery, many schemes are proposed based on traditional cryptographic methods, such as homomorphic and commutative encryption, which inevitably introduce tremendous overhead to the system and are not practical for the resourcelimited mobile devices. In this paper, we propose an Efficient Weight-based Private Matching (EWPM) protocol, which provides a realistic matching approach considering both the number of common interests and the corresponding weights on them. In EWPM, we utilize the Confusion Matrix Transformation (CMT) algorithm to achieve a secure and efficient matching. Security analysis and detailed simulations show that our proposed scheme can realize privacy-preserving friend discovery with higher efficiency.

A Differentially Private Matching Scheme for Pairing Similar Users of Proximity Based Social Networking applications

“…Various solutions have been proposed in recent years [1,7,8,9,13,14] to address the issues of preserving privacy while matching users. These solutions all assume the user has multiple interests chosen from a public set of defined interests.…”

“…Different from [10] our work does not rely on third-party matching or computationally expensive encryption algorithms. Also, we ensure differential privacy for users data, which [8] and [14] suffer from as well. Our results show relatively accurate matching, compared to the ambiguous results of [8].…”

Efficient private matching based on blind signature for proximity-based mobile social networks

“…mul 1 , mul 2 , exp 1 and exp 2 denote one 1024-bit multiplication, 2048-bit multiplication, 1024-bit exponentiation, and 2048bit exponentiation, respectively. We assume that the public attribute set in our scheme has n interests, each user has n p attributes, and n c (n c ≤ n p )common attributes and the highest corresponding weight value is l. As shown in Table-I, we compare our scheme with Fine-grained [15], WAS [16], EWPM Level − I privacy [18], EWPM Level − II privacy [18] and two-party scheme of [17]. From Table-I, it is obvious that our scheme significantly reduces computation and communication costs, especially in terms of the online computation cost which directly effects system performance.…”

“…Unfortunately, too much communication and computation overhead is introduced in their paper. Considering the tremendous overhead caused by traditional cryptographic methods, Zhu et al [18] proposed an efficient confusion matrix transformation algorithm to achieve secure and efficient matching. In their scheme, however, users can only order the attributes and the corresponding attribute values for each attribute in the public set to construct the confusion matrix, regardless of whether they have attribute or attribute value, which causes extra communication overhead.…”

A Comment on Privacy-Preserving Scalar Product Protocols as Proposed in “SPOC”

“…As shown in [2], the protocol is much faster than publickey based protocols using homomorphic encryption. Since then, this protocol has been and is still used in many privacy-preserving solutions, e.g., [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], including support vector machines [17], facial expression classification [9], medical pre-diagnosis [18], and speaker verification [10], [11].…”