Also covering Zero-fee commitments

V3 commitments are LN commitment transactions made using version 3 transactions adhering to the policies for TRUC, a P2A outputs, ephemeral dust, and sibling replacement. This allows it to be a zero-fee commitment, as the onchain fee will be paid by a CPFP spend of the ephemeral P2A output. The commitment transaction and its fee-paying child will be transmitted by package relay.

LN commitment transactions commit to the current state of the channel, including each party’s settled balance plus any pending HTLCs that conditionally transfer funds between them. Commitment transactions usually aren’t published, allowing for offchain balance updates, but each transaction must be confirmable onchain in case one party needs to close the channel without the cooperation of the other party. For a transaction to be confirmable, it must be both valid and incentivize miners to include it in a block—which is usually done by having it pay a transaction fee.

The original LN-Penalty commitment transaction format used entirely endogenous fees, with the full fee amount being implicitly allocated in the transaction itself using normal Bitcoin transaction semantics. This had the advantage of being simple and compact, but it was fragile if market pricing for feerates suddenly spiked—a previously signed valid commitment transaction might no longer be confirmable in a reasonable amount of time, which can lead to loss of money.

An updated commitment format called anchor outputs or anchor commitments added two additional outputs to the commitment transaction to allow either party to CPFP fee bump it if necessary. This allowed reducing the commitment transaction fee to the minimum necessary to get it to relay. Although this was a major improvement, it was still fragile if the minimum relay feerate increased. Additionally, it depended on CPFP carve-out policy, which is vulnerable to transaction pinning attacks and may be removed to allow deployment of cluster mempool.

Since the initial proposals in 2018 for anchor commitments, Bitcoin Core transaction relay policy and tools have seen significant advances. V3 commitments build on those improvements:

  • ● TRUC (v3) transactions significantly reduce the worst-case pinning vulnerabilities. Additionally, CPFP carve-out required that only the two outputs designed for fee bumping be immediately spendable, forcing all other outputs (such as HTLCs) to include a 1-block relative locktime (commonly called “1 OP_CSV”). This prevented using any of the value in those outputs to pay transaction fees.

  • ● Package relay allows the commitment transaction to pay zero fee. This eliminates the fragility of increases in feerates (or minimum relay feerates) making a previously signed commitment transaction non-confirmable. It also allows removing the update_fee mechanism from the LN protocol, which was a source of multiple problems, from stuck channels to security vulnerabilities.

  • ● P2A allows the output intended for fee bumping to be spendable by either party. Additionally, P2A uses only a small amount of onchain space. Sibling RBF allows either party to increase the fee bump even if the other party created the initial fee bump.

  • ● Ephemeral dust allows the P2A output to have a zero value even though it would otherwise be uneconomic. This is allowed under the assumption that spending a small P2A output is the most efficient way to add fees to the zero-fee commitment transaction using public transaction relay.

Primary code and documentation

Optech newsletter and website mentions

2025

2024

2022

See also

Previous Topic:
Version 2 P2P transport

Next Topic:
Vaults

Edit page
Report Issue