superchain-backend design doc

[protolambda]

Description

Design for the superchain-backend, the next iteration of the idea after initial takes by @hamdiallam in Q1.

This supports the cross-L2 message verification for superchain interop.

[protolambda]

Comments summary:

• Optimization of acquiring data via RPC, rather via P2P or some aggregate or optimized form.
• Protocol change to be able to verify initiating/executing message inclusion more easily (i.e. not as inefficient as full receipt proofs)
• Flexible trust model, not everyone may want to run this new service.
• Need to expand on redundancy and reliability. May need load-balancer to be aware of unsafe vs cross-unsafe.
• One remaining comment from Hamdi that we did not get to during the design review call.

Doc to be updated, and can then do a final review/merge.

[protolambda]

Updates on previous review comments and requests:

• Optimization: added optimization ideas to throughput sub-section in considered risks.
• Protocol change for message-inclusion proofs: listed as option under light-client idea, but out-of-scope for now.
• Flexible trust model: design-doc already offers a general interface and some optionality, I believe this forms sufficient forward-compatibility with other trust models.
• Redundancy and reliability: added a new section covering this, with proposal how to use multiple backends as rollup node.
• Comment from Hamdi about block-safety resolution: added missing details on implementation summary, nice catch.

Additionally:

• Naming: we are calling the standalone service the op-supervisor, since it monitors superchain activity for cross-L2 safety.
• Rebased on main
• Moved the design-doc to protocol/ directory.