Verification of cryptocurrency consensus protocols: reenterable colored Petri net model design

Abstract

The reliability of consensus protocols establishes the basis for cryptocurrency cybersecurity. Formal and simulation techniques are appreciated for the verification of proof-of-work agreement protocols, because the formal techniques development process frequently requires a series of assumptions, leading to somewhat unrealistic models. A conventional colored Petri net model, presented by the authors in a previous study, allowed us to refine the Keller and Böhme consensus protocol by applying the simulation technique. In this study, we developed a reenterable colored Petri net model representing a general construct for model-driven development and refinement of consensus protocols and the corresponding software. A considerable benefit of a reenterable model is the invariance of its structure with respect to the net topology, number of attached nodes, and parameters of their software and hardware, which are represented as the marking of dedicated places. Switching of topology tags associated with dynamic objects represents the basic principle of reenterable model functioning. It is shown that, having the same characteristics, in bounds of admissible error of some 2-3%, the reenterable model results in 6 times speed-up of the simulation process and a drastic decrease in time for model editing, especially when the number of nodes exceeds a hundred. Thus, a reenterable model is a valuable complement to possible formal models that allows us to obtain express-evaluations of characteristics in a rather short time and with an acceptable level of adequacy.