Actually same question goes into all L2 (godwoken, axon ) protocols and L3 dApps
Godwoken transaction data (other than account information and wrapped assets) is stored in the witness data of CKB (similar to calldata in Ethereum)
Axon transaction data is stored completely off-chain, only headers and some information needed to run consensus is stored on CKB
Wondering why we need two L2s? Would it not of been best to have updated Godwoken?
Enlightening, thanks Matt!
The biggest issue I see is the difficulties in implementing challenges/fraud proofs in Godwoken and all of the problem space around decentralizing sequencers. Checking a fraud proof for EVM->CKB-VM is a different universe from EVM->EVM, it requires alot of R&D.
Without these (as well as alternative data availability solutions to increase transaction capacity), Godwoken loses in the comparison to other EVM L2’s.
Additionally, leveraging CKB’s account abstraction is really the only path to any sort of differentiated offering to begin with.
The Axon team has been working heavily on making CKB-VM work with Axon (compared to Godwoken, which has had comparatively little work in this respect) and Axon offers more mature and well understood security mechanisms (proof of authority/proof of stake)
I agree with you it would be ideal if Godwoken accommodated all of the motivation in an optimistic rollup, but the effort to do so would severely take away from other priorities (such as L1 development) and simply put, Axon is a more convenient target.
Imo what is missing is a team seeing promise in Godwoken as a generalized solution, which can offer account abstraction and support any VM, not just EVM. The addition of outside effort (and funding) would allow the core team to continue on their priorities for L1 and see the potential of Godwoken continue to be realized.
Does “off-chain storage” mean 3rd parties like arweave…?
it would be stored by the Axon validators, so it’s “on-chain” in the context of the Axon chain but “off-chain” in the context of CKB