Decentralized Asset Custody Scheme with Security against Rational Adversary

TL;DR

This paper introduces a decentralized asset custody scheme supporting many custodians and assets, ensuring security against rational adversaries while reducing overhead and improving liveness in digital asset management.

Contribution

It proposes a novel combinatorial design-based decentralized custody framework with security guarantees against rational adversaries, supporting large-scale asset management.

Findings

Supports large number of custodians and assets

Reduces overhead cost through small custodian groups

Guarantees security against rational adversaries

Abstract

Asset custody is a core financial service in which the custodian holds in-safekeeping assets on behalf of the client. Although traditional custody service is typically endorsed by centralized authorities, decentralized custody scheme has become technically feasible since the emergence of digital assets, and furthermore, it is greatly needed by new applications such as blockchain and DeFi (Decentralized Finance). In this work, we propose a framework of decentralized asset custody scheme that is able to support a large number of custodians and safely hold customer assets of multiple times the value of the total security deposit. The proposed custody scheme distributes custodians and assets into many custodian groups via combinatorial designs, where each group fully controls the assigned assets. Since every custodian group is small, the overhead cost is significantly reduced. The liveness is also improved because even a single alive group would be able to process transactions. The security of this custody scheme is guaranteed under the rational adversary model, such that any adversary corrupting a bounded fraction of custodians cannot move assets more than the security deposit paid. We further analyze the security and performance of our constructions from both theoretical and experimental sides and give explicit examples with concrete numbers and figures for a better understanding of our results.