Investigating Blockchain Mining Games Using a Simulation Framework

Abstract

As blockchain technologies have grown in popularity, many have considered the security properties of the longest chain consensus protocol used by Bitcoin and other popular blockchain implementations. As done by Kiayias et al., Ferreira and Weinberg, and others, we model blockchain mining under longest chain consensus as a stochastic game. Instead of taking an analytic approach, as many others have done, we use simulation data to assess the space of possible mining strategies.

In order to obtain empirical data about different mining games, we introduce mining-sim, a simulation library implemented in Rust which can execute a blockchain mining game consisting of tens of millions of blocks on the order of a few seconds. Using this simulation framework, we replicate results from the literature and explore the caveats of implementing mining strategies as algorithms.