Bitfold

Blog

Ethereum: The Merge. Moving consensus to PoS

Posted: 07 September 2022

photo: Miloslav Hamřík / Pixabay

Ethereum: The Merge. Moving consensus to PoS

Ethereum is the first and most used smart-contract enabled blockchain. Since its creator was a Bitcoin developer, its consensus mechanism relies on a proof-of-work (PoW) model. However, its upcoming and long-awaited update “The Merge” changes it significantly. Key elements of this technological update will include the alteration of ETH consensus mechanism from PoW to proof-of-stake (PoS) model. The goal of this article is to explain the difference between those two very different technical approaches, the implications of The Merge, and some potential network-wide attack vectors for both.

Why PoW works

PoW is the consensus mechanism introduced by Satoshi in the Bitcoin network. It allows any party to be part of the network, without any limitations - the only need is to run the current version of software, to have a network connection and proper equipment to start hashing.

As miners get a subsidy from creating new blocks as well as all transaction fees included in it, there is an economic incentive to work alongside other miners instead of trying to attack the network. Any miner that produces a block containing an invalid transaction, can announce it - but the rest of the network nodes will drop it, even the not-mining ones. By doing so malicious miners simply waste their hashing power instead of receiving a reward.

How PoW is safe

At this point, it is important to understand the difference between the computational cost of creating a block and its verification. Miners need to do additional PoW calculations to create blocks (the current network-wide Ethereum PoW hashrate is about 900TH/s = 900*10^12 hashes per second), but the rest of the nodes just validate all that work by checking a single hash. This translates to the very high security of the current network setup.

The only known attack possibility that is capable of harming a PoW-based blockchain is the so-called 51% attack when malicious miners own over half of network-wide hashing power. In this scenario, attackers can rewrite part of the blockchain and “swap” the content of a transaction. However, the current hash rate of the whole Ethereum network is far from being owned by one party, as the amount of mining equipment and energy needed is enormous; the double-spend transaction needs to be in millions of dollars to make an attack even remotely profitable.

Pros and cons of PoS

PoS works on different principles. Instead of hashing power block producers need to have the proper amount of coins for a given blockchain. That way, the mining process is more energy-efficient, as it does not need hashing. The biggest downside of the PoS mechanism is that one can not simply join a network - they need to possess a determined minimal amount of coins - 32ETH. Given the current price of 1 coin (approx. $1500), the ability to become a validator is limited.

Another common PoS problem is that - as there is no “punishment” mechanism for malicious block producers - it may lead to constant forks in the network - as block producers can deliver many different blocks at the same height without any cost. This can lead to a situation, where any transaction can be treated as confirmed, as it can be included in a block in one fork but not in others. Every blockchain should be easily verifiable at a given moment, if we have concurrent forks of the same height, there should be a way to determine which one to pick. In PoW model – the fork with a higher hashrate confirms it is the longest. There is no certainty like that in the PoS model.

The change that The Merge brings

The Merge is a process of joining an already working PoS beacon chain (sometimes called ETH2) with PoW Ethereum chain, and then removing PoW from Ethereum. This way users that already own and use Ethereum do not need to do anything - as there are no changes in the execution layer (EVM state itself) nor account balances.

Currently, operating miners become obsolete, and only properly configured full nodes operated by owners of locked 32ETH coretellar are needed to run the network. After-merge full node configuration will require running two separate clients - one for the consensus layer (beacon chain) and one for the execution layer (EVM). As execution nodes need reports to beacon one hash of the state to be notarized, also receiving new blocks on the execution state requires validation of the previous state hash to be sure that EVM state is intact and in sync with other nodes.

To be continued

Currently, the Merge is not going to affect network speed or transaction fees - as change is exclusively on the consensus layer.

Only when sharding is enabled in future network upgrades will this lead to raised network capacity and speed. Therefore those who are already expecting a significant change, especially regarding the environmental aspects, should still be patient.