One of the many contradictions associated with blockchain technology is that for all it proselytizes about the benefits of decentralization and freedom from third-party interference, centralization vectors keep appearing. 

One area where this has become particularly concerning is the Execution Client software used for the Ethereum blockchain, which currently has approximately 70% of nodes using Geth. This supermajority issue is a genuine concern for Ethereum stakers and the community as a whole, should there be a consensus failure within a supermajority client, then that could lead to a reorganization of the chain. If, for instance, Execution Client A accepted a reorganized version of the chain, that wouldn’t be a problem if the other Execution Clients B, C, and D say it’s wrong; however, if there’s a supermajority of over 66% confirming this, then there’s a real crisis. 

Not entirely a theoretical case

Although this may seem theoretical, it’s not. On Jan. 21 this year, a bug in Ethereum’s Nethermind client software—used by blockchain validators to interact with the network— knocked out a chunk of the chain’s key operators. Given that Nethermind powers approximately 8% of Ethereum’s validators, the situation was manageable, and the Ethereum blockchain continued to operate normally.

Should Geth be afflicted with a bug, this could fatally undermine the Ethereum blockchain. There are two break points for proof-of-stake (PoS) blockchains when one-third disagree and two-thirds disagree. Should over a third disagree, the Ethereum blockchain would slow down, and validators would face an inactivity penalty; however, the situation would still be manageable. If, however, more than two-thirds of validators sign a block and vote for it, and it becomes part of the chain, there would be a finalized but invalid chain. In that situation, the Ethereum community would be forced to fork.

This issue has added poignancy with the potential approval of Ethereum-based exchange-traded funds (ETFs) later this year, should institutional investors use a supermajority Execution Client for staking ETH and there’s an error when attesting to the blockchain, the potential problem could see stakers with large sums of ETH lose everything in a short space of time.

There are currently 28,976,695 ETH at stake on the network. Approximately 70% of this (approximately 20 million ETH) can be attributed to validators running Geth, and 16% (around 5 million ETH) to validators not running Geth. For the non-Geth chain to finalize, the validators running Geth must have their stake burned until it represents less than one-third of the remaining total stake. It means around 21.5 million ETH would need to be burned from these validators (around 90% of their stake), reducing the Geth stake to approximately 2.5 million ETH, representing less than one-third of the 7.5 million total ETH at stake (2.5 million plus 5 million ETH). The 5 million ETH controlled by the non-Geth validators would now represent over two-thirds of the stake, allowing them to finalize the chain. This would be an excruciatingly painful process that would play out over approximately 40 days. It would be so significant that it would reduce the total supply of all ETH by around 18%, bringing the total supply below 100 million ETH. The potential consequences of errors occurring within Geth are too much for current stakers to bear.

There is more at stake 

It’s also important to acknowledge that this isn’t a problem that only Ethereum has to deal with; other PoS blockchains have this genuine issue of Execution Client software dependency. Ethereum is, however, unique in that the community is at least attempting to ensure greater diversity in its Validator Client software, preventing this from becoming an issue.

Digital currencies were born out of the aftermath of the 2008 financial crash, which saw governments bail out banks that were deemed “too big to fail,” it would be a cruel irony if an industry that aimed to avoid the errors of traditional finance merely ended up replicating them. Few people predicted the financial crisis then; however, people are beginning to wake up to the potential problems over-reliance on Geth can lead to. Unlike then, the danger is clear to see, and the solution is at hand; as a community, we should look to seize the moment and ensure Validator Client software is monitored and maintained to ensure diversity within the ecosystem is guaranteed for the greater good of those involved.

Steve Berryman is a chief business officer and co-founder of Attestant. With a Ph.D. in Computer Science, Dr. Berryman has over 25 years of software engineering experience within the financial industry. He has worked for tier-one investment banks across the globe, where his primary focus has been on cutting-edge derivatives and software to price and manage their risk. Steve is responsible for running one of the largest Ethereum-validating infrastructures in the world and is working with institutional financial firms to make Ethereum staking more accessible to everyone.