Overview
Blockchain protocols commonly operate in open networks where message delay is difficult to regulate. Traditional blockchain protocols typically assume synchrony, whereas more recent protocols aim to maintain security despite unexpected network delays. Within the traditional model, which involves $n$ participants categorized as either honest or Byzantine, a supermajority of $\lfloor 2n/3\rfloor + 1$ honest participants is generally required for security.
Research has investigated reducing the number of required honest nodes within a game theory model. This reduction is sometimes achieved to a simple majority by incorporating $k$ rational players and $t$ Byzantine players. The presented research introduces STAKE (Secure and Tolerant Algorithm through ($k,t$)-robust Equilibrium), a blockchain protocol designed to function effectively with only $\lfloor n/3\rfloor + 1$ honest players.
Research Context
Conventional blockchain designs often necessitate a high proportion of honest participants, specifically a supermajority of $\lfloor 2n/3\rfloor + 1$, to prevent malicious activities like double spending. This requirement stems from models where $n$ participants are classified strictly as either honest or Byzantine. Efforts in game theory models have explored scenarios where a mix of rational and Byzantine players can lower the honest node requirement to a simple majority.
The challenge has been to reduce this honest node threshold further while maintaining security against colluding majorities, particularly regarding double spending attacks. The STAKE protocol addresses this by proposing a mechanism that allows for a significantly lower proportion of honest participants than previously established in such models.
Approach
The STAKE protocol mandates that each consensus participant stakes a specific amount, $s$, which must be sufficiently large in comparison to their total liquidity, $\ell$. This staking mechanism is central to its operational model.
The protocol's efficacy is analyzed in the context of $\zeta$-uple double spending attacks, which refer to an attacker's ability to execute $\zeta$ such attacks. The methodology involves demonstrating how the probability of a coalition successfully executing such attacks decreases. This analysis supports the claim that STAKE establishes a $(k,t)$-robust equilibrium.
Findings
The STAKE protocol successfully operates with a reduced requirement of $\lfloor n/3\rfloor + 1$ honest participants, a lower threshold compared to the $\lfloor 2n/3\rfloor + 1$ supermajority typically needed in traditional blockchain models.
A key finding is that the probability of a coalition managing to execute $\zeta$ double spending attacks, referred to as a $\zeta$-uple attack, diminishes polynomially fast with increasing $\zeta$. This characteristic is crucial for the protocol's security guarantees.
Through this mechanism, STAKE achieves a $(k,t)$-robust equilibrium. This equilibrium signifies that rational players are disincentivized from colluding with Byzantine players. This outcome is attained even with a relatively low ratio of the staked amount to liquidity, $s/\ell$.
Why This Matters
The ability of STAKE to function securely with only $\lfloor n/3\rfloor + 1$ honest players signifies a potential shift in the resource requirements for blockchain network security. Reducing the honest participant threshold could lower the barrier to entry or operational costs for maintaining decentralized networks. The demonstration of a $(k,t)$-robust equilibrium implies that rational actors are unlikely to collaborate with malicious entities, bolstering the protocol's resilience against collusion, even with modest staking requirements.