ライブラリー > Full Analysis of Nakamoto Consensus in Bounded-Delay Networks
March/2020, ePrint Archive
Nakamoto consensus, arguably the most exciting development in distributed computing in the last few years, is in a sense a recasting of the traditional state-machine-replication problem in an unauthenticated setting, where furthermore parties come and go without warning. The protocol relies on a cryptographic primitive known as proof of work (PoW) which is used to throttle message passing with the PoW difficulty level being adjusted appropriately throughout the course of the protocol execution.
While the original formulation was only accompanied by rudimentary analysis, significant and steady progress has been made in abstracting out the protocol’s properties and providing a formal analysis under various restrictions, starting with the work by Garay, Kiayias and Leonardos [Eurocrypt ’15], for a simplified version of the protocol which excluded PoW difficulty adjustment and assumed a fixed number of parties as well as synchronous communication rounds. These assumptions have since been somewhat relaxed, first by Pass, Seeman and Shelat [Eurocrypt ’17] who also focused on the simplified version of the protocol but on the bounded-delay model of communication, and by Garay, Kiayias and Leonardos [Crypto ’17] who looked into the full protocol including the PoW difficulty adjustment mechanism with a variable number of parties but assuming synchronous communication and a predetermined schedule of participation. Despite the above progress, the full analysis of the protocol in the more realistic setting of bounded delays and dynamic participation has remained elusive.
This paper’s main result is the proof that Nakamoto’s protocol achieves, under suitable conditions, consistency and liveness in bounded-delay networks with adaptive (as opposed to predetermined) dynamic participation assuming, as before, that the majority of the computational power favors the honest parties. While our techniques draw from previous analyses, our objective is significantly more challenging, demanding the introduction of new techniques and insights in order to realize it.