Proof-of-work is a consensus mechanism that many blockchains use to secure their ledgers. In it, workers (called “miners”) compete with one another to solve a mathematical equation. First to solve “wins” the block, its coinbase (the reward for winning), and the associated fees from transactions in the block.
However, when hashing power becomes too centralized, a network can become susceptible to what is known as a “51% attack.” When someone controls at least 51% of a network, they may be able to successfully reorganize transactions and commit “double spends,” much like we saw on the Ethereum Classic network in January.
Yet with every 51% attack of top 100 networks we’ve seen, none of them have gone away completely – and in cases like Bitcoin Gold and Ethereum Classic, we didn’t even see too much sway in price, market cap, or hashing power. Why?
I have a few opinions on this, and while I am in no way suggesting that network hash security is unimportant, I don’t believe infrequent 51% attacks to be existential threats to a network.
A successful attack would require:
- A network worth attacking
- At least one victim – often an exchange with good liquidity for the attacked token
- Secrecy in both identity of the attacker(s) and ultimate destination of the double-spent or cashed-out coins.
A network worth attacking typically means one that already has strong community support. While frequent 51% attacks would absolutely corrode the trust a community has in its token, infrequent ones are unlikely to do all that much to shake this.
A victim worth scamming, in most cases exchanges, also would likely become privy to such an attack rather quickly. These exchanges also typically require KYC and AML identification of its users which leads to the next point.
There’s no point in stealing if you’ll get caught. Apart from privacy coins, tracking transactions on chain (while laborious) is 1) possible and 2) provable.
However, cryptocurrency users should still aim to make it even harder for would-be attackers by introducing more miners and more hashes into various chains. If the frequency of such attacks determines their ability to kill a network, it is important to make them as infrequent as possible.
Ryan Condron is CEO of Titan Industries.