Hostile Blockchain Takeovers (Short Paper)
Most research modelling Bitcoin-style decentralised consensus protocols has assumed profit-motivated participants. Complementary to this analysis, we revisit the notion of attackers with an extrinsic motivation to disrupt the consensus process (Goldfinger
- PDF / 232,225 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 14 Downloads / 232 Views
ntroduction
Bitcoin [15] has achieved significant popularity since its 2009 launch, with a monetary base nominally worth over US$100 billion at the time of this writing. Perhaps Bitcoin’s most important innovation is its decentralised consensus protocol. Bitcoin-style consensus (or “Nakamoto consensus”) uses computational (proofof-work) puzzles to maintain consensus on the blockchain, a public append-only ledger storing all transactions to prevent double-spending. The computational puzzles are intended to make disrupting the consensus protocol expensive, as an attacker must obtain a large fraction of all computational power in the system to deviate from the default protocol. This basic design has been adapted in dozens of follow-up cryptocurrencies with similar consensus protocols, notably Ethereum [21] which is itself worth close to US$30 billion. It was known from the start that an attacker with a majority of computational power can easily cause arbitrarily deep forks in the blockchain [15]. It has subsequently been shown that an attacker with substantially less power can, at the very least, undermine the fair distribution of rewards in the system [4,16,18]. These attack strategies are profitable in a fixed exchange-rate model (in which an attacker’s utility is solely measured in currency units within the system itself). A similar modelling approach has been used in many papers [6–8,10,11,17] analyzing Bitcoin-style protocols with the goal of proving positive results about incentive-compatibility; that is, that given a specific utility model for miners intended properties of the system will emerge such as an ever-growing longest chain (stability) and proportional distribution of mining rewards (fairness). An inherent limitation to this approach is that real-world attacks may negatively affect systems’ value (and exchange rate with external currencies), making some mining strategies which deviate from the standard protocol to increase c International Financial Cryptography Association 2019 A. Zohar et al. (Eds.): FC 2018 Workshops, LNCS 10958, pp. 92–100, 2019. https://doi.org/10.1007/978-3-662-58820-8_7
Hostile Blockchain Takeovers
93
nominal miner revenue actually yield less utility. A more realistic utility function is revenue denominated in a stable external currency (such as US dollars). Because accurately modeling the impact of miner behavior on exchange rates is difficult, analysis along these lines is usually qualitative. Thus, our ability to compare the stability of competing protocol flavours like proof-of-work and proof-of-stake remains limited. In this work, we analyse the stability of consensus protocols from a different viewpoint. Rather than considering the risk of an attacker undermining desired properties to maximise utility, we consider an attacker whose explicit goal is to undermine and destabilise the consensus protocol. Kroll et al. [11] first considered such an attacker, which they called a Goldfinger attacker after the James Bond villain who attempted to irradiate US Treasury reserves. This attack
Data Loading...
