Match & Privacy: Bridging Bidirectional Discovery in Mobile Social Networks

TL;DR

Social networking often involves a paradox: we want to find "like-minded" people based on specific traits, but revealing those traits to strangers is a privacy nightmare. This paper proposes a breakthrough bidirectional matching scheme. Unlike traditional models where you just "unlock" someone else's data, this system requires a mutual policy match verified by a cloud intermediary—without the cloud or the users seeing the underlying attributes or failing criteria.

The "Passive Participant" Problem

Current mobile social networks (MSNs) usually utilize Ciphertext-Policy Attribute-Based Encryption (CP-ABE). In these systems, a sender encrypts a message with a policy (e.g., "Age > 20 AND Hobby = Hiking"). The receiver is passive; if they have the keys, they open it.

The authors identify a critical gap: Receiver Autonomy. Why should the receiver be forced to accept a match just because they fit the sender's criteria? True social matching should be a two-way street—a mutual "opt-in" where both parties' requirements are met simultaneously before any private data is revealed.

Methodology: The Outsourced "Double-Handshake"

The core innovation lies in combining Matchmaking Encryption with Decryption Outsourcing. The architecture involves three main entities: the Sender, the Receiver, and a Semi-Trusted Outsourcing Center (OC).

1. The Dual-Policy Construction

Each party defines an LSSS (Linear Secret-Sharing Scheme) access structure. The matching succeeds only if:

• Sender's Attributes satisfy the Receiver's Policy.
• Receiver's Attributes satisfy the Sender's Policy.

2. Efficiency through Outsourcing

Bilinear pairings are the "expensive" parts of these cryptographic protocols. Instead of making a smartphone do the heavy lifting, the users generate Transformation Keys (TK). These keys allow the OC to partially decrypt the ciphertext.

Figure 1: The process involves the OC performing time-consuming transformations, returning only a simple, constant-size ciphertext to the user.

The magic here is that the OC performs the work but stays blinded. It only knows if $1+1=2$ (a match occurred); it cannot see the attributes ($x$) or the policies ($P$).

Experimental Validation: Breaking the Linear Growth

In standard ABE, as you add more requirements (hobbies, location, profession), the time to match grows linearly. On a mobile phone, this eventually becomes a battery-killer.

Performance Highlights:

• Constant Local Cost: The local decryption time remains constant even as the number of attributes increases.
• Ciphertext Compression: The OC transforms complex LSSS ciphertexts into short El Gamal ciphertexts.

Table 1: Comparison showing the proposed scheme's support for bidirectional matching and reduced local computational ops compared to SOTA.

Figure 4: Standard ABE decryption time skyrockets with attribute complexity, whereas the proposed method stays nearly flat for the end-user.

Critical Analysis & Future Outlook

The scheme is a massive win for mobile efficiency. However, the authors rely on a "semi-trusted" cloud. If the OC were truly malicious, it might attempt to return false negatives.

Key Takeaways for Future Research:

• Verifiability: Future iterations need a way for the user to verify the OC actually performed the transformation correctly (Verifiable Outsourced Decryption).
• Fully Untrusted Clouds: Moving toward protocols that assume the cloud is actively trying to cheat would be the next logical step for high-security applications.

In summary, by shifting the heavy lifting of attribute-based matching to the cloud while maintaining a bidirectional "consent" gate, this paper paves the way for a more private and efficient era of mobile digital discovery.