Abstract
Identity-Based Restrictive Partially Blind Signatures (ID-RPBS) are critical cryptographic primitives designed to balance user privacy with regulatory and systemic constraints in applications like electronic cash. Recently, Wang, Tang, and Li proposed an ID-RPBS scheme claimed to be secure against forgeries. This paper presents a rigorous cryptanalysis of that scheme, demonstrating a fundamental flaw in its algebraic structure. We show that an adversary can maliciously alter the non-blinded, agreed-upon information within a signature—such as a monetary denomination—with a 100% success rate. This vulnerability allows for the creation of valid signatures on forged data, effectively rendering the Wang-Tang-Li (WTL) scheme insecure for practical use.
Introduction
In the landscape of digital finance, electronic cash systems must satisfy several competing requirements. They must ensure user privacy (blindness), allow for the inclusion of public metadata like expiration dates or denominations (partial blindness), and enforce strict formatting to prevent double-spending (restrictiveness). Identity-Based (ID-based) cryptography further simplifies these systems by allowing public keys to be derived from known identifiers, such as email addresses.
The ID-RPBS scheme proposed by Wang, Tang, and Li (WTL) was intended to provide these features securely using bilinear pairings and the Computational Diffie-Hellman Problem (CDHP). However, a closer examination of the protocol's interaction phase reveals a critical weakness. This paper details how the WTL scheme fails to achieve existential unforgeability, allowing a malicious user to bypass the "restrictive" and "partially blind" safeguards of the protocol.
The ID-RPBS Primitive
To understand the failure of the WTL scheme, one must consider the three pillars of an ID-RPBS:
- Blindness: The signer (typically a bank) issues a signature on a message without learning the message's content. This protects the user's privacy during transactions.
- Partial Blindness: Both the signer and the user agree on a piece of common information (
info) that remains visible and un-blinded. This is used for administrative data like currency value. - Restrictiveness: The message must conform to a specific structure. This prevents users from embedding unauthorized data into the blinded message.
If any of these properties are compromised, the integrity of the electronic cash system is lost. Specifically, if the info parameter can be altered by the user after the signer has processed the request, the user can unilaterally change the value of a digital coin.
Cryptanalysis of the WTL Scheme
The vulnerability identified in the WTL scheme stems from the mathematical handling of the info parameter. The protocol utilizes a hash function, $H_1$, to incorporate the common information into the signature challenge.
The Cancellation Flaw
The core of the attack lies in the linear nature of the interaction between the signer and the user. During the signing phase, the user receives a commitment from the signer and responds with a blinded challenge. The authors demonstrate that because the signer’s response is a linear combination involving the hash of the info, a malicious user can calculate a mathematical offset.
By carefully crafting this offset using the properties of the underlying group, the user can "cancel out" the hash of the original agreed-upon information (e.g., "$100") and substitute it with the hash of forged information (e.g., "$100,000,000"). This manipulation occurs entirely on the user's side during the blinding process, and the resulting signature remains mathematically consistent with the forged information.
Results and Impact
The cryptanalysis yields several definitive conclusions regarding the WTL protocol:
- Probability of Success: The attack is not probabilistic or dependent on brute force. An adversary following the described algebraic steps can successfully forge a signature with a probability of 1.
- The "Million Dollar" Forgery: To illustrate the severity, consider a scenario where a bank agrees to sign a digital note with
info = {2008.01.01, $100}`. Using this attack, the user can produce a valid signature for `info' = {2010.12.31, $100,000,000}. - Verification Success: Because the forgery exploits the internal logic of the signature generation, the resulting forged signature passes all standard verification equations. A merchant or a secondary bank would have no mathematical reason to reject the forged $100 million bill, as it appears to be a legitimate product of the signer’s private key.
Conclusion
The findings of this paper provide a formal proof that the Wang-Tang-Li ID-RPBS scheme is insecure against existential adaptive chosen-message-and-identity attacks. The vulnerability highlights a recurring danger in cryptographic design: the risk of linear dependencies in interaction phases.
For future development of ID-based restrictive partially blind signatures, it is clear that the "common information" must be more robustly bound to the algebraic structure of the signature. This paper serves as a reminder that even schemes built on hard problems like the CDHP can be undermined by subtle flaws in protocol logic. Designers must ensure that no part of the agreed-upon metadata can be isolated or cancelled out by a malicious participant during the blinding process.
Summary updated