RFC: Handle large storage proofs.

[rahulksnv]

1. When the storage proof in InvalidStateTransitionProof exceeds a threshold, a summary of the storage proof is included in the fraud proof instead of the full storage proof.
2. The verifier is expected to recreate the storage proof locally and verify that the summary is correct.

The PR is meant to get feedback on the approach/feasibility for now, will be opened for further review based on feedback.

Issue: #1596

Code contributor checklist:

• I have read, understood and followed contributing guide

[vedhavyas]

Would have been nice to open the draft on your fork and requested feedback there to avoid PR cluttering.

What we want to achive here is similar to Ethereum's OutOfGas. In our case, since we do not execute extrinsic through gas metering, we want to introduce a metering on Storage reads where we will have max amount of storage read bytes and mark a transaction as failed if the read bytes exceeds the max value. Max value for each transaction will be setup such that a fraud proof carrying such reads in the Storage proof will fit in the consensus block. If a malicious operator ends up not exiting when such read data is exceeded and submits an ER, any honest node can produce invalid state transition fraud proof and verifier can run this transaction and comes to same error as honest node there by verifying the invalid state transition.

Hope that make sense

[rahulksnv]

What we want to achive here is similar to Ethereum's OutOfGas. In our case, since we do not execute extrinsic through gas metering, we want to introduce a metering on Storage reads where we will have max amount of storage read bytes and mark a transaction as failed if the read bytes exceeds the max value. Max value for each transaction will be setup such that a fraud proof carrying such reads in the Storage proof will fit in the consensus block. If a malicious operator ends up not exiting when such read data is exceeded and submits an ER, any honest node can produce invalid state transition fraud proof and verifier can run this transaction and comes to same error as honest node there by verifying the invalid state transition.

I did spend two days looking at adding OutOfGas kind of functionality on the upstream run time execution environment. That certainly is a bigger effort, may be more than a week(also mentioned here: paritytech/substrate#14350). This PR was meant to get feedback on alternate(less perfect) proposal.

If OutOfSpace error during execution is what we want, I can go back and work on that. @jfrank-summit @vedhavyas please confirm