Evaluating fork after withholding (FAW) attack in Bitcoin

Abstract

Fork after withholding (FAW) attack is an easy-to-conduct attack in the Bitcoin system and it is hard to be detected than some attacks like selfish mining and selfholding attacks. The previous studies about FAW attack made some strong assumptions, such as no propagation delay in the network. This paper aims to quantitatively examine the profitability of FAW attack in Bitcoin system with block propagation delay. We first establish a novel analytic model, which can analyze FAW attack in the Bitcoin system. Then we apply the model to design metric formulas for the Bitcoin system. These formulas can be used to evaluate the miner profitability (in terms of miner reward) and the impact of FAW attack on system throughput (in terms of transactions per second). We make a comparison of FAW attack and other attacks (including selfish mining and selfholding attacks). Experimental results reveal that FAW adversaries can get more rewards in the network with propagation delay than without delay. The results of the comparison of selfish mining and FAW attacks show that adversaries with large computational power can conduct selfish mining or selfholding attack to get more rewards, but they can conduct FAW attack to profit more when their computational power is small. Our work can be used to analyze Bitcoin-like blockchain systems and help design and evaluate security mechanisms.