An important property of blockchains is consistency, which is a guarantee that all honest parties output the same sequence of blocks throughout the execution of the protocol. Prior analyses of consistency have either made some simplifying assumptions about network synchrony, or have required a careful counting of combinatorial events that is not easy to do for many protocol variants. In this talk, we will present a simple Markov-chain based method for analyzing consistency properties of blockchain protocols. The method includes a formal way of stating strong concentration bounds as well as easy ways to concretely compute the bounds. We will illustrate the use of the framework to obtain tighter consistency guarantees, study new protocol parameter regimes, analyze new attack families, and provide a rigorous consistency analysis of recently proposed protocols.

Video Recording