Securing Proof-of-Stake Blockchain Protocols

Abstract

Proof-of-Stake (PoS) protocols have been actively researched for the past five years. PoS finds direct applicability in open blockchain platforms and has been seen as a strong candidate to replace the largely inefficient Proof of Work mechanism that is currently plugged in most existing open blockchains. Although a number of PoS variants have been proposed, these protocols suffer from a number of security shortcomings; for instance, most existing PoS variants suffer from the nothing at stake and the long range attacks which considerably degrade security in the blockchain.

In this paper, we address these problems and we propose two PoS protocols that allow validators to generate at most one block at any given “height”—thus alleviating the problem of nothing at stake and preventing attackers from compromising accounts to mount long range attacks. Our first protocol leverages a dedicated digital signature scheme that reveals the identity of the validator if the validator attempts to work on multiple blocks at the same height. On the other hand, our second protocol leverages existing pervasive Trusted Execution Environments (TEEs) to limit the block generation requests by any given validator to a maximum of one at a given height. We analyze the security of our proposals and evaluate their performance by means of implementation; our evaluation results show that our proposals introduce tolerable overhead in the block generation and validation process when compared to existing PoS protocols.