Ethereum in the PoS (Proof of Stake) era is getting harder and harder to understand. The newly introduced staking system has spawned what seems like a never ending stream of hard to grasp gotcha’s. As I noted, the more Ethereum succeeds, the worse things seem to get.
This month Christine Kim reports that the biggest problem is the number of validators [2].
Ethereum’s current design invites anyone to bid to be a validator, requiring a stake of 32ETH (you can big more than once, in increments of 32). The idea is that, when you participate you can get some of the fees, and make a profit validating blocks for the Ethereum blockchain.
This seems to be an attractive proposition, because there are currently hundreds of thousands of validators, heading toward millions. Obviously, this is a good thing, as far as assuring that the network doesn’t become dependent on a small set of nodes. And, if I understand correctly, it probably helps finance the network.
But, as noted earlier, this ties up all those 32ETH stakes, e.g., 500,000 validators would mean that 16 million ETH are out of circulation.
And, as Kim notes, at these scales, a peer-to-peer network starts to break down through sheer volume. Ethereum hasn’t crashed yet, but if the trend in growth of validators continued, nodes will be overwhelmed just with messages between validators .
The peril of this trend is reflected by a sudden consensus among Ethereum developers, and proposed rapid changes to cap this growth [1]. Crypto developers never change their minds, and never agree with each other. So a sudden burst of mind-changing rush to agreement is very striking.
I’m not going to make pronouncements about engineering the Ethereum protocol. My own instincts tend to be highly skeptical of the ability of Carbon-based Life Forms to make gigantic decentralized protocols work. So, on general principles, I’m not optimistic about solving the problems. Each “solution” will bring new problems.
I will note that this kind of issue is not exactly surprising. The reason that most large scale networks are not peer-to-peer is that P2P doesn’t scale up. No matter what you do, the bigger the network, the worse the performance of the protocols. It’s literally just math.
My second comment is to note that this experience is reminding us why the Great Emperor Nakamoto went with a Proof of Work protocol. PoW is (a) simple and (b) scales well with the number of validators. The classic Bitcoin PoW is a blunt instrument, which wastes unholy amounts of energy. But, like any blunt instrument, it is easy to understand and has few moving parts to go wrong.
Sure, Bitcoin has scaling problems, but they are really simple problems. Too many users, too much data.
I’m not a huge fan of PoW protocols, but Ethereum is showing us why you would want to keep things simple (and blunt).
- Christine Kim (2023) The Most Pressing Issue on Ethereum is Validator Size Growth. Coindesk, https://www.coindesk.com/consensus-magazine/2023/09/29/the-most-pressing-issue-on-ethereum-is-validator-size-growth/
- Christine Kim, Paths Toward Reducing Validator Set Size Growth. Galaxy, 2023. https://www.galaxy.com/insights/research/paths-toward-reducing-validator-set-size-growth/
Cryptocurrency Thursday
Robert McGrath's Blog 
One thought on “Yet More Ethereum Protocol Issues”