catchup: state-proof-based catchup #5720
Draft
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add support for catchup to use state proofs to validate new blocks.
The main changes are:
The catchup service now has a state proof fetcher, whose job is to retrieve state proofs for rounds beyond the current ledger state. These state proofs are stored in an sqlite DB, since there might be many state proofs that we need to fetch.
The catchup service exposes a way to set trusted "renaissance" parameters for authenticating the initial state proofs, for cases when we can't use state proofs from the genesis block (like the situation we have on mainnet now).
The BlockService HTTP interface adds support for retrieving a state proof, and for getting a light block header proof instead of a cert when retrieving a block.
The catchup service uses state proofs, if possible, to authenticate new blocks, in lieu of agreement certificates. The catchup service is backwards-compatible: if it requests a state proof from the BlockService, but receives an agreement certificate instead (e.g., because the BlockService has not been upgraded with the above changes), the catchup service will validate the certificate instead.
The config file has additional fields to optionally specify the renaissance catchup parameters, to allow catchup to start validating state proofs from some (trusted through out-of-band channels) block.
It may be a good idea, for performance, to pre-compute the state proofs and distribute them in a single file, rather than asking many relays to find the state proofs on-demand. This is not done yet, because it largely depends on how we would want to distribute these bundled state proofs. It should be reasonably straightforward to feed a bundle of state proofs into the stateProofFetcher.
This is a somewhat large PR; I tried to factor out some independent pieces from it (the still-open ones are PRs #5663, #5703, #5718), but unfortunately I couldn't find other substantial pieces to meaningfully split out on their own. This PR is against a temporary branch (
nickolai/state-proof-catchup-base
) that already has the above PRs applied, so that they don't show up in the code diff for this PR.