Is centralisation the real Ethereum-killer?
by Hugh Brooks, Director of Security Operations at CertiK
Decentralisation is a core tenet of blockchain technology. In fact, decentralisation is blockchain’s raison d’être; there’s simply no need for a distributed ledger if all users of that ledger are fine with centralisation’s conveniences and compromises.
Blockchain technology excels in potentially adversarial environments where a single point of failure or manipulation would be unacceptable.
Yet centralisation is not a binary, it’s a spectrum. And decentralisation is a goal that must be of paramount importance if blockchain technology is to differentiate itself from centralised systems and continue to promote financial inclusion for users around the world.
Since Bitcoin’s quasi-immaculate conception over a decade ago, the network has experienced waves of centralisation, decentralisation, and recentralisation.
Initially, there was only one miner: Satoshi. Then, as the cryptocurrency gained popularity, thousands of amateur enthusiasts used their home computers to contribute to the network. Now, the vast majority of hash power is concentrated in a handful of entities which operate vast mining farms full of highly specialised equipment.
Ethereum – the second most valuable blockchain in the world – was born in 2014 out of a well-publicised ICO rather than being gifted to the world by a still-pseudonymous founder. It has never laid claim to the same standard of decentralisation as Bitcoin: almost one-fifth of all ETH was reserved for the Ethereum Foundation and the project’s founders at the time of the ICO.
Proponents of proof-of-stake (PoS) have argued that Ethereum’s transition away from proof-of-work (PoW) – which has secured the Bitcoin blockchain since its inception – will increase the network’s level of decentralisation. These claims are worth examining. The stakes are nothing less than the survival of Ethereum as a free and open platform.
Claim #1: Staking makes it easier for individuals to participate in securing the network, promoting decentralisation. A validator node can be run on a normal laptop. Staking pools allow users to stake without having 32 ETH.
Staking requires no special hardware, unlike proof-of-work which necessitates expensive and energy-hungry mining rigs. All staking requires is just that: a stake in the network. In theory, staking removes the costly barriers to entry that gatekeep mining in 2022. However, in practice most people don’t have the interest or expertise to run their own nodes. Instead, they delegate their Ether to staking providers, who add the ETH to their node and pay out (most of) the rewards earned.
The result, as we’ve seen post-’Merge’, is that centralised staking providers control just as large a proportion of the network validating power as Bitcoin miners do.
And geography matters, too. 45% of Ethereum validator nodes are located in the United States. This is worryingly close to a majority of all nodes, which could become a real problem for the health of the network if US authorities were to impose strict(er) compliance requirements on node operators.
It doesn’t take too large a leap of faith to picture a scenario where the US government mandates the censorship of transactions associated with hostile countries under the guise of anti-money laundering or counter-terrorist financing. Such action would undoubtedly have the effect of catching innocent users – or even internal dissenters to these hostile regimes with no other open financial tools available to them – in the crossfire.
Already, approximately 25% of all blocks created since the Merge utilise MEV-boost, which increases profits for validators while conforming to OFAC sanctions.
Claim #2: Staking is more decentralised. Economies of scale do not apply in the same way that they do for PoW mining.
Broadly speaking, PoS removes many of the layers of abstraction that exist in PoW mining. The economies of scale associated with energy costs and access to hardware and capital do not apply. Instead, PoS introduces a different sort of economy of scale: the more you own of the network’s native asset, the more of it you receive as block rewards. The rich get richer in a very linear relationship that does not exist in PoW mining, where miners on the Bitcoin network are not necessarily required to maintain any exposure to BTC itself.
The argument goes that this relationship between ETH owned and ETH earned incentivises validators to protect the integrity of the network. This is true, but there are other considerations to take into account. If the majority of staked ETH is delegated to a handful of staking pools, and these pools take a cut of the rewards earned (which are paid out in ETH), then the tendency is for the operators of these major staking pools to grow disproportionately richer and more powerful.
Overall, the PoS vs PoW debate is a complicated and sometimes heated one. Ethereum’s implementation of a PoS consensus mechanism is just weeks old and it will take time for the full impact of the Merge to be realised. While Proof of Stake has drastically reduced Ethereum’s energy consumption, it has not had such radical effects on the level of network centralisation.
There is a real threat that Ethereum (and other blockchains) will continue to tend towards centralisation. Beyond a certain point there is little chance of re-decentralising the network. Now is the time to prioritise decentralisation to ensure that free and fair technology continues to improve the lives of all people around the world.