State Root Reasoning on L1 #305
Replies: 1 comment 9 replies
-
Regarding the necessity of posting state roots on L1, it's crucial for ensuring that the state transitions on the OP Mainnet occur correctly according to the EVM. While Ethereum guarantees the correctness of the world state transitions through its consensus algorithm, Gasper, OP Mainnet being a rollup solution relies on Ethereum to provide this assurance. If state roots weren't submitted to L1, there wouldn't be a mechanism to ensure that the state transitions on OP Mainnet are properly validated according to the EVM rules.
You can visit: |
Beta Was this translation helpful? Give feedback.
-
|
In bitcoin, it's addressed by proof of work, but if majority collude maliciously, you got the same problem in bitcoin. As far as I understand, you base your logic on a fact that majority nodes wouldn't collude on bitcoin because it already has so many nodes, whereas on OP mainnet, it's easier to collude as there won't be those many nodes. Q1: Is this what you meant ? If so, OP Mainnet's whole idea is to not depend on L2 nodes. Basically, If I want to know the balance of my account, first, I ask L2 node to give me the balance and the state trie proof and I use the root hash from L1 chain for that tree to verify. This ensures that even if ALL L2 nodes are malicious, that won't make anything horrible/broken as long as there's a single verifier existing. Q2: All correct ? |
Beta Was this translation helpful? Give feedback.
-
|
In the context of Bitcoin implementing Nakamoto consensus, external entities can trust the network as a whole if they believe that over 51% of it is honest. However, without consensus mechanism, external parties have no way to confirm if state transitions are correct unless they participate in the network by running a full node. Your method implies reliance on a single verifier. Indeed, in a situation where a specific individual maintains the entire infrastructure of an Optimistic rollup and is the sole user of that rollup, there might be no need to submit state roots to L1. This is because the trust and verification would be concentrated in the hands of that single participant. |
Beta Was this translation helpful? Give feedback.
-
|
Yes, I think I was correct in my last reply. In bitcoin, if 51% is honest, then at some point, longest chain rule will apply and make the state of the network correct. In the OP mainnet case, we dont have longest chain rule and we should design the system such that even if 99% of L2 nodes are malicious, our system still should succeed. Even if only 1 verifier out of the whole OP mainnet nodes is honest, with L1 roots, system still work. Though, when a user(external entity) asks for a balance to L2 node, that user cant trust it directly the same way It would on L1. So I think the flow is this: user would get the latest state root from L1 and then ask the L2 node for state proof and then use that state ROOT to verify the proof. I think thats my final observation and I am sure you exactly agree with that, just because of my non-full trust in myself, I will ask again: am I correct ? 😂 |
Beta Was this translation helpful? Give feedback.
-
In such a case, while I believe the network would become practically unusable, the system would still function as designed. Discussing what happens in this scenario honestly goes beyond my understanding. You might find this blog post helpful: https://vitalik.eth.limo/general/2020/08/20/trust.html
Strictly speaking, you would indeed need to use the state root from L1 for verification. However, in reality, users who do not operate a full node are essentially placing their trust in a single full node, whether it's on L1 or L2. |
Beta Was this translation helpful? Give feedback.
-
|
Sure... L1 state roots seem to exist for the following: "let it be here, we won't need it, but in case there's a need, we must have the central, correct authority to resolve our problem". I will think more tomorrow and will let you know if I have cleared my doubts and mark the answer accepted. Huge, limitless thanks to you for bearing with me all along. <3 <3 |
Beta Was this translation helpful? Give feedback.
-
Did you check the documentation?
Did you read the specs?
Did you check for duplicate questions?
Issue Description
My question will be related to optimism v0 that you guys were building 3 years ago. Currently, it mightn't be the same as my question will relate to, but I'd appreciate the answer as if you were answering optimism v0(ofc, if you know, if not, then you can answer it with the latest optimism in mind)
As we know, sequencer publishes calldatas and state roots to L1. The reason it publishes calldatas is that it should be super easy to recover L2 chain in case something goes wrong. Since ethereum is secure, we rely on it so that in case Optimism goes down, we can still recover the chain by re-executing calldatas.
Now, I wonder why state roots are published. One reason you will say is that it allows fraud proofs, such that verifier can check if the state root published is correct and if not, prove on L1 why it's not correct, but I believe this is not the reason why we need state roots published on L1. Even if you don't publish it, L2 nodes still exist, they still wait for incoming blocks from sequencer and in case, tx is not valid in a block, L2 nodes won't add it to their chain. Everything will still be great.
So, then I researched and read that, Its purpose is to allow secure withdrawals.
I am curious how this works. When you want to deposit on L2, you call L1's smart contract's
depositfunction which emits the event to which L2 node listens to and runs it on L2's EVM. All clear now, but it seems like state roots on L1 is needed to withdraw funds from L2 to L1.Thank you so much.
Additional Information
No response
Feedback
No response
Beta Was this translation helpful? Give feedback.
All reactions