📝 SummarXiv

Abstract

Traditional security models for Nakamoto-style blockchains overestimate adversarial coordination by assuming instantaneous synchronization among malicious nodes, neglecting the critical impact of internal communication delays on security. This paper introduces a dual-delay framework to revisit security analysis, addressing this oversight through two key innovations. First, the static delay model quantifies how adversarial communication delays (\(\Delta_a\)) constrain the effective growth rate of private chains, derived via an M/D/1 queuing model as \(\lambda_{eff} = \lambda_a / (1 + \lambda_a \Delta_a)\). This model reveals that the security threshold (\(\beta^*\)), the maximum adversarial power the system tolerates, increases with \(\Delta_a\), even exceeding the classic 51\% boundary when \(\Delta_a \textgreater \Delta\) (honest nodes' delay), breaking the long-standing 50\% assumption. Second, the dynamic delay model integrates probabilistic corruption and scale-dependent delays to characterize the total adversarial delay window (\(\Delta_{total} = \Delta(n) e^{-k\beta} + c \log(1 + \beta n)\)), where \(\Delta(n) \in \Theta(\log n)\) captures honest nodes' logarithmic delay growth. Asymptotic analysis shows adversarial power decays linearly with network scale, ensuring the probability of \(\beta \leq \beta^*\) approaches 1 as \(n \to \infty\). By exposing the interplay between network scale, communication delays, and power dilution, we provide a theoretical foundation for optimizing consensus protocols and assessing robustness in large-scale Nakamoto-style blockchains.