Recap of Metrics on the Eve of the Merge
Folks, the moment we have all been waiting for has arrived! The Merge is finally happening on Ethereum, slated for September 14–15 at time of writing.
As we prepare to head into the new world of Proof of Stake Ethereum, we recap key metrics related to staking.
Metrics looked at include:
- Deposit amount
- Validator distribution
- Node distribution
- Client distribution
Deposit amount
The Beacon chain deposit contract is now by far the largest single holder of ETH in the entire Ethereum network, with 11.3% of all ETH staked in it.
After gradually dropping off over 2021, deposits into the contract shot up in March 2022 with its highest ever monthly amount. It retained the momentum to mid-May, at which point it slowed down again.
Grouping the data by entity shows that the additional deposits came almost entirely from Lido, boosted by key integrations such as that with Aave. These catapulted Lido to the top of validators. As they approached 33% of total stake in May, Lido self-limited their deposits — June onwards saw minimal deposits from them.
Validator distribution
Naturally, Lido’s quick capturing of stake share led to concerns about the decentralization of Ethereum. At first glance, it appears as if Lido alone makes up almost 1/3 of stake and that only 3 entities make up more than 50% of Ethereum validators.
However, Lido is made up of 29 separate node operators. Splitting up the data across each node operator makes for a less concentrated chart. In this view, 12 entities are required to make up 50% of validators.
How does this compare to Proof of Work Ethereum? In PoW Ethereum, three mining pools alone make up more than 50% of hashrate.
Since OFAC added Tornado Cash to their list of sanctioned addresses, talk of potential censorship has been rife within the Ethereum community. Some were concerned that the Ethereum protocol itself would be censored if the validators were to be compelled by authorities in their jurisdictions.
Let’s look at the breakdown of depositors by jurisdiction. Inferring the jurisdiction of the top 50 known entities from their company registrations, we find that slightly less than 1/3 of validators are within a single jurisdiction (U.S.).
One thing to note is that Coinbase, which makes up almost half the U.S. portion, has made a commitment to shut down staking rather than censoring transactions if the situation calls for it.
Node distribution
A similar concern with regards to centralization is on the distribution of nodes. Three separate data sources show that the U.S. and Germany make up more than 50% of all nodes.
When it comes to hosted nodes, the data is less consistent. Eth2 Nodewatch reports 75% of nodes being non-hosted while Ethernodes reports almost the exact opposite — 70% hosted nodes.
Digging deeper into Ethernodes’ hosted node numbers, we see that a vast majority of these are based in the U.S. and Germany, which tracks with the geographical data above.
Staking service providers may want to consider diversifying their node distribution to ensure better stability for themselves and the network!
Client distribution
Another important aspect of distribution is client diversity for both the consensus and execution layers. Having too many nodes running a particular client risks a liveness failure for the entire chain and penalties for ALL the validators running that client if that client encounters a bug.
On this point, the consensus layer distribution has greatly improved. Several months ago, Prysm made up 75% of all consensus clients. Today, Client Diversity estimates it to be only 30–43%. In an ideal world, no single client should have more than 1/3 of stake.
Execution client diversity remains a major challenge. While clients such as Erigon and Nethermind have been improving, Geth remains far and away the client that most nodes rely on. In case of a critical bug with Geth, all of the validators that use it for the execution layer would face penalties to their stake.
As seen in this article, the robustness of Ethereum’s decentralization is a work in progress. Above and beyond these decentralization metrics, a final defense layer of the Ethereum network is its community. Should an attack on the chain occur and the metrics above prove insufficient to protect against it, the community can get behind social slashing to fork Ethereum, remove the stake of the attacking party, and carry on with our lives.
What are your thoughts on the metrics and the Merge? Let us know!
Credit to Bernard Lim for data analysis.
