Original Title: Endgame
Original Author: Vitalik Buterin
Original Link: https://vitalik.eth.limo/general/2021/12/06/endgame.html
The central idea of the full text:
- No matter you are a friend of Ethereum or a killer, join the Rollup camp, let’s roll together!
- From a long-term perspective, the centralization of Ethereum Rollup block production may be an inevitable trend, But, That's OK...
Join the RollUp camp
Current "big blockchains" (BSC, Solana, etc.) have very high block frequencies, very large block sizes, thousands of transactions per second, and are highly centralized. To put it simply, ordinary servers cannot run nodes.
Generally, dozens or even hundreds of nodes can run a fully participating block. How can this be "self-proving innocence", after all, such a block looks relatively "centralized".
You can learn Ethereum, introduce RollUp, OP fraud proof or ZK cryptography proof, etc.
specific method:
- Add a second layer of collateral. The transactions in the block are divided into multiple "buckets", and each secondary pledger is randomly assigned to one of the buckets to verify the state root of the Merkle tree, and 2/3 verifiers in each bucket sign the next block Only then will it be accepted, and the matter of block verification will be "decentralized".
- Apply OP (fraud proof) or ZK-Snark like Ethereum.
- Introduce data availability sampling. Allowing users to use a light client to randomly select some blocks for verification is equivalent to giving the people the right to randomly check whether you have done anything wrong.
- Add secondary transaction channels to prevent censorship. So, what kind of BSC or SOL (or any other big blockchain) do we get when these methods are finally used?
- Their block-producing nodes are still very centralized; however, block verification is trustless and highly decentralized; special anti-censorship can prevent block producers from conducting censorship.
- Aesthetically uglier, but works well. Even if every block producer intends to attack or censor, the worst case scenario is to go offline completely, the chain stops trading, until the community pools resources and establishes an honest main stakeholder node.
The focus here is to decentralize the use of verification. For TPS, that is, to accommodate more users and experience, centralization of block generation is inevitable, so decentralized verification of these decentralized nodes.
Ethereum RollUp is inevitable
The block production (block production) of Ethereum Rollup expansion may also gradually become centralized.
There are two reasons:
1. One situation is that a certain RollUp is doing well.
Whether it is Arbitrum, Optimism, Zksync or any new generation of Rollup, as long as the hardware is strong enough, the TPS can reach 10,000 per second, the cost is low, and the experience is very good. After Layer2, everyone finally ran to this Rollup chain (essentially, this became a Rollup Layer2 in a parallel world to Ethereum Layer1).
In Layer 2, block production is centralized, and block verification is trustless and highly decentralized.
In Rollup, the hardware requirements of the sequencer node are very high . Many Rollups only use an official centralized sequencer from the beginning.
But using a network of centralized sequencers only compromises trust and censorship resistance, not security.
And the RollUp sequencer operator cannot steal funds.
2. Another possibility is that no RollUp is the best, and everyone chooses competition + cooperation
The Ethereum Rollup eventually becomes a multi-chain style similar to Cosmos. Everyone does different things on each Rollup, shares the security of ETH Layer1, and then jumps around on each Rollup through various bridges.
There is a high probability that this will happen:
- Cross-domain MEV (miners can extract value) will be discovered by nodes, and it is more beneficial to do several Rollup nodes at the same time, such as Eden, when multiple Rollups coexist, there are only a lot of opportunities for this kind of cross-domain MEV, under economic incentives , there will be nodes that are multiple Rollup nodes at the same time.
- Regardless of the economic incentive of cross-domain MEV, in terms of technology, a good Optimism block producer is likely to be a good Arbitrum block producer, a good Polygon block producer, and even a good Ethereum base layer block producer. This kind of resource and technology will have a gathering nature. In the end, those nodes with the strongest technical strength and the best reputation will serve as the nodes of each Rollup, with a high degree of overlap.
In the end, the big blockchain, only one RollUp survives, and multiple RollUps compete and cooperate. These three routes all lead to the same final result: block production is centralized, but block verification is trustless and highly decentralized. And still protect against censorship.
in conclusion
Don't rush to predict which RollUp will win, keep an open mind. According to the consistent style of ETH, such an end will take several years, and cross-domain MEV research has just started, but the road is tortuous, and the future is still bright!