Scaling Lightning Safely With Feerate-Dependent Timelocks

The email from Antoine addresses the concept of a feerate-dependent timelock (FDT) proposal for Bitcoin, which aims to provide a solution for blockchain congestion detection, particularly in the context of Lightning Network and smart contracts. The FDT proposal is introduced as a novel formalization that utilizes sliding-time-locks based on block fee rates.

Antoine explains that the FDT mechanism prevents the inclusion of a transaction within a block under certain conditions. These conditions are determined by existing consensus rules, such as bip68 and bip113, along with the implementation of new parameters like 'feerate_value_bound,' 'window_size,' and 'block_count.' These parameters allow transaction builders to set thresholds for when a transaction's confirmation can be delayed based on the median fee rate of a defined number of blocks.

For instance, in a Lightning Network penalty scenario between two parties, transactions can be pre-signed with a specified 'feerate_value_bound.' Antoine provides an example where if the median fee rate surpasses this bound within a selected window, the transaction would not be included in the blockchain. This mechanism is suggested as a potential solution to the issue of "Forced Expiration Spam" described in the Lightning Network paper.

Antoine also identifies design considerations for implementing FDTs, including the modification of revocable scripts to inspect FDT parameters and the addition of a 'claim_grace_period' parameter to mitigate fee-rate races after revocation timelocks expire. He suggests that all channels carrying a Hash Time-Locked Contract (HTLC) should have consistent FDT parameters to prevent issues like off-chain HTLC double-spending.

The recommendation is made to incorporate FDT parameters into the bip341 annex due to available space and the ability to set varying FDT parameters for individual HTLC outputs within a single transaction. This could facilitate future covenant mechanisms like HTLC aggregation.

While Antoine does not claim that FDTs make timeout-tree protocol more efficient than existing channel factories and payment pool constructions, he expresses optimism about the proposal's potential to address vulnerabilities related to blockchain congestion in time-sensitive applications. However, he notes that the analysis and game theory aspects concerning miner collusion have not been thoroughly reviewed, especially considering the proposed introduction of the 'claim_grace_period.'

The email concludes with Antoine expressing enthusiasm for the progress represented by the FDT proposal in addressing some pressing issues in the Lightning Network and other similar use cases.