Public.icon

This is of course a problem not only for distributed oracles but also for the Bitcoin protocol itself (e.g., Goldfinger attack).

https://gyazo.com/a16cd4c0d2411c2f397bfef5d977660c

Let's introduce the scenario of a Goldfinger attack on Bitcoin. You are the largest Bitcoin mining farm. Mining is not easy. In addition to hardworking and depreciating ASICs, Moore's Law is a headwind, new players are entering the field, and the relationship with the government is delicate and constantly changing (we're watching you, China). And finally, as the price of BTC/local currency softens and your margin disappears in the flames of a red candlestick, there is a deep and liquid derivatives market for BTC. How about stacking cheap put or short futures contracts, doubling down on trades, spamming nodes, and moving hash power to other SHA-256 protocols? If you have political influence in the community, propose a frivolous fork with complicit developers to undermine trust. Basically, it's a series of attacks on an unreliable payment system.

Goldfinger attack aims to crash the value of the targeted cryptocurrency by using the majority of voting rights within the system to weaken the underlying consensus protocol. In the context of Proof-of-Stake, which is based on the amount of cryptocurrency held by the voting power, it can be achieved through a buy-out attack that buys over 50% of the target currency.

In this context, the RTTD (Race to the Door) effect is explained, which causes more holders to exit before the target currency becomes worthless. This effect causes the price of shares with voting rights to decrease further and become cheaper as the attack progresses. This paper aims to demonstrate the technical feasibility of Race to the Door style attacks without acquiring a majority of voting rights (i.e., hash rate) in the context of Proof of Work (PoW). To achieve this goal, we examine the technical feasibility and cost of such attacks using Ethereum as an example. First, we present a system model for RTTD attacks on PoW-based cryptocurrencies and explain the overview. This divides the attack into preparation, race, and attack phases. To demonstrate technical feasibility, these phases are implemented in the form of smart contracts on Ethereum. In the attack phase, three variations are presented, each of which achieves a different type of denial-of-service attack. Band payments are used to incentivize either the trigger of additional transactions or the creation of empty blocks by miners. To estimate the cost of the proposed attack variations, we conducted empirical analysis by examining transaction data from past congestion periods on the Ethereum blockchain. Based on the results, we estimated and compared the costs of attack variations. The cost per hour of blocking transactions by inducing further transactions is about 870 Ether. The cost of incentivizing miners to empty one-third of the blocks per hour is about 790 Ether. This work demonstrates that Race to the Door attacks are technically feasible in the context of PoW-based cryptocurrencies. The cost of the attack depends on its strength and duration. The strength can be set in the attack phase, and the duration depends on the amount available for the attack.