How to Beat Nakamoto in the Race

TL;DR

This paper introduces a Markov decision process framework to analyze and quantify the most effective attack on proof-of-work blockchain safety under network delays, providing precise violation probabilities.

Contribution

It develops an MDP model for blockchain security analysis, proposes an optimal attack strategy, and calculates exact safety violation probabilities under various conditions.

Findings

Identifies bait-and-switch as the optimal attack strategy.

Provides exact probability calculations for safety violations.

Offers new insights into network delay effects on blockchain security.

Abstract

This paper studies proof-of-work Nakamoto consensus protocols under bounded network delays, settling two long-standing questions in blockchain security: What is the most effective attack on block safety under a given block confirmation latency? And what is the resulting probability of safety violation? A Markov decision process (MDP) framework is introduced to precisely characterize the system state (including the blocktree and timings of all blocks mined), the adversary's potential actions, and the state transitions due to the adversarial action and the random block arrival processes. An optimal attack, called bait-and-switch, is proposed and proved to maximize the adversary's chance of violating block safety by "beating Nakamoto in the race". The exact probability of this violation is calculated for any given confirmation depth using Markov chain analysis, offering fresh insights into the interplay of network delay, confirmation rules, and blockchain security.