Briefing
The foundational problem of long-term key security in distributed systems is the mobile adversary model, where a persistent attacker compromises different subsets of parties over time, eventually accumulating enough secret shares to reconstruct the key. This research introduces a provably secure proactive secret sharing protocol that solves the practical constraint requiring all $n$ parties to be online for a refresh cycle. The breakthrough is achieved by folding the share refresh computation directly into the standard threshold signature generation process, using the resulting on-chain signature as a public commitment and signal. This new theory’s single most important implication is the enabling of truly resilient, long-term distributed custody solutions that maintain proactive security for real-world devices that are frequently offline.
Context
Traditional threshold signature schemes were designed for a static security model, assuming the adversary could only compromise a fixed number of parties $t$ at any single point in time. Proactive Secret Sharing (PSS) was developed to counter the more realistic mobile adversary, requiring all parties to periodically engage in a complex, multi-round, interactive protocol to refresh their shares. This “all-online” requirement created a critical theoretical and practical limitation, especially for consumer-grade threshold wallets where devices are often powered off or disconnected, thereby compromising the system’s long-term security and practical liveness.
Analysis
The core mechanism is the integration of the share refresh protocol into the signature generation itself, a concept termed “folding.” When a required threshold of online parties collaborate to create a signature, they simultaneously generate a new set of secret shares for the next epoch. The resulting final signature, which is broadcast and recorded on the blockchain, contains cryptographic elements that act as a publicly verifiable commitment to the new shared secret. Offline parties, upon waking, can observe this public on-chain signature, verify its validity against the old public key, and then use the committed data within the signature to non-interactively compute their new, updated secret share. This design fundamentally differs from prior approaches by decoupling the refresh protocol’s liveness from the availability of all $n$ participants, leveraging the blockchain as an asynchronous, public bulletin board for cryptographic state updates.
Parameters
- Offline Support → Any number of parties $n-t’$ can remain offline during a refresh cycle, provided $t’$ online parties meet the signing threshold $t$.
- Communication Rounds → Zero extra rounds are required for the online signing parties, as the refresh is folded into the standard signature protocol.
- Compatibility → The scheme is compatible with standard threshold versions of ECDSA, EdDSA, and Schnorr signatures.
- Adversary Model → Achieves security against a mobile adversary, where the attacker can compromise different subsets of parties over time.
Outlook
This foundational work immediately enables the development of next-generation distributed key management systems, particularly for multi-party computation (MPC) wallets and decentralized autonomous organizations (DAOs). In the next three to five years, this asynchronous security model will become the standard for high-security digital asset custody, allowing for robust, long-term secure solutions that do not sacrifice user experience or device autonomy. The research opens new avenues for exploring how on-chain state can be leveraged as an information-theoretic signal to coordinate complex, asynchronous cryptographic protocols across a vast network of intermittently connected devices.
