Note: The original text comes from @yorkzhong’s long tweet.
1. Components of a modular blockchain:
- Execution layer: Rollup/L2/L3, performs the calculations required by users/smart contracts, and is understood as a computer
- DA: All original transactions at the execution layer are ensured to be stored here for later verification and confirmation, which can be understood as a database
- Settlement: Verification/Confirmation/Finality. Users can raise objections here, and the original execution layer data stored in DA is required.
There is also a consensus layer, which mainly contains the contents of Layer 1 and will not be elaborated here.
2. Three levels of main technology stacks or representative products:
The executive layer mainly includes:
- ON STACK
- ZK Stack
- Cosmos Stack
There are three types of settlement:
- Eth layer1
- Dymension (settlement in cooperation with Celestia)
- Layer2 itself (e.g. Arbitrum, which can be used as a settlement for the next-level Layer3)
There are about 5 types of DA:
- ETH blob/ Danksharding
- Celestia (Independent decentralized DA)
- Near D.A
- EigenDA/MantleDA (DA using ETH restaking)
- Centralized DA (Validum Mode)
Figure 1 is a comparison of different DAs made by Near. Ethereum still has an ultimate state of danksharding that is not included here, which is reasonable because Ethereum has not yet determined the timetable for danksharding (Cancun upgrade is 4844, prototype Danksharding).
From a technical perspective, Ethereum Danksharding and Celestia are the most decentralized because they both use sampling technology, which reduces the performance requirements of nodes while achieving large bandwidth.
EigenDA comes second and also uses sampling, but EigenDA is parasitic on Ethereum and its number of nodes is a subset of Ethereum.
Others probably don’t use sampling, for example, the degree of decentralization of NearDA is equivalent to that of Near Protocol.
3. Modular blockchain combination:
The above creates approximately 3 * 3 * 5 = 45 possible combinations, for example:
- Arbitrum: OP (execution layer technology stack) + ETH L1 (settlement layer) + ETH blob (DA)
- Mantle: OP+ETH L1+MantleDA
- XAI: OP + Arbitrum L2 + third-party DA
A hundred flowers bloom
4. Security differences between different modes
What are the differences between these different combinations? In addition to the differences in fees and TPS, the Ethereum system will of course maintain the "orthodoxy" of Ethereum. For example, Vitalik believes that if Ethereum's DA is not used, then Ethereum cannot guarantee the safe withdrawal of assets. He calls all solutions other than Ethereum DA "Validum". (Figure 2)
Of course, this is just one person’s opinion, but using the “short board theory”, for example, if you settle on Ethereum, but the DA uses another company (such as Near), then what you actually get is Near-level security because the minimum security value is taken.
From another perspective, one of Celestia's concepts is to focus on settlement, DA, and social layer (ha, community). For example, if you are an RU on Celestia and the project owner does something bad, the "community" can choose to use the DA on Celestia to do a hard fork and completely reconstruct the chain. This is also an angle, but it sounds like nuclear deterrence.
-----------------------------------------------------------
The following is my personal speculation and is for reference only:
5. Future Settlement Layers The future settlement layers will be very rare, with about 2-3 major settlement layers. There are several issues involved:
- Cross-chain asset transfer across settlement layers will still be a challenge, but cross-chain asset transfer within the same settlement layer is relatively easy to achieve securely.
- The more complicated issue is the interoperability. First of all, interoperability is an inevitable step in the future. Web2 has very strong interoperability. It is hard to imagine that Web3 will not be interoperable in the future. That would be a huge step backwards. Interoperability requires the same settlement layer. Different settlement layers cannot interoperate because finality and protocols cannot be coordinated. Cosmos' IBC also mainly emphasizes interoperability within the IBC system.
- In addition, the decentralization cost and economic security requirements of the settlement layer are relatively high, so there will only be about 2-3. Whether BitNet can become one of the settlement layers remains to be seen. At present, it is understood that Bitcoin's UTXO model is not yet sufficient to support settlement, but it may be possible after future upgrades.
6. Future DA layer
DA layers have sprung up like mushrooms after rain. The core content is that DA is not that difficult. The simplest one can do DA on a single machine, and the most complex one is the decentralization under sampling like Celestia.
By the way, the more nodes a sampling model uses for decentralization, the greater the bandwidth (imagine a p2p movie download network), which also has a network effect, so I don’t think there will be too many such “decentralized sampling DAs” in the end.
But other forms of DA have no limit and can be infinite.
Here we can use an analogy, DA is storage, and the cost is very high. If you do not strictly require Ethernet-level security services, choosing which DA is a trade-off between cost and security.
The principle is: the more valuable the service, the safer the DA should be.
Therefore, the DA in the future may be decentralized. Even so, perhaps 7-8 major DAs may be enough.
7. Narrative impact on ether (personal speculation)
If the DA layer in the future is very fragmented, wouldn’t that be bad for Ethereum?
The barriers to entry for DA are indeed not very high. Ethereum DA has its advantages, but it does have the problem of high prices. However, Ethereum DA can still eat up the most high-end and most profitable part of the market like Apple.
The future of Ethereum should be L2 interoperability, that is, the layer 2 on Ethereum can interoperate with each other. This should be difficult for other public chains to forcibly copy.
Can interoperability capture value? Of course it can. In fact, the ATOM 2.0 white paper has already given some ideas, and it is not a big problem for Ethereum to learn from them again.
In short, if you have value first, it will never be difficult to capture it.
