LNPBP-53: Multipeer channels

LNPBP: 0053
Layer: OSI Application (i7)
Vertical: Lightning network protocol
Title: Muli-peer payment channels for Bifrost
Author: Dr Maxim Orlovsky <orlovsky@lnp-bp.org>
Comments-URI: https://github.com/LNP-BP/lnpbps/pulls/97
Status: Draft
Type: Standards Track
Created: 2021-11-14
Finalized: not yet
License: CC0-1.0

Abstract

This specification proposes a batch of improvements for lightning payment channels: (1) generalization of channels to have an arbitrary number of participants (multi-peer channels), (2) abstraction from specific peyment routing syncrhronization algorithm (HTLC, PTLC), (3) ability to extend and change channel parameters and structure on-the-go, (4) use of Schnorr signatures, MuSig, Taproot and miniscript-based descriptors. It achieves this properties by leveraging Bifrost protocol, specifically LNPBP-50 and LNPBP-51 standards.

Background

Lightning network enables scalable and more private bitcoin payments by using pre-signed off-chain transactions, composing so-called state channels. Still, existing lightning network, defined by a set of BOLT standards, has multiple limitations. LNP/BP standards LNPBP-50, 51, 52 proposed an extended version of the lightning network called Bifrost. It is a part of more general Lightning network protocol and allows creation of more complex, functional and secure forms of payment channels, which is the aim of this proposal.

Motivation

Current standards defining lightning network (BOLTs) were created in the eraly SegWit era. They were created to cover single use case (unilaterally funded dual party payment channels) with a very restricted functionality (no channel splits, channel factories, nested channels, extensions for channel structure etc). Since LN launch a multiple enhancemend had been deployed on Bitcoin maiinnet, including Shnorr signatures, unlocking more compact multisigs (with MuSig scheme), scriptless scripts & adaptor signatures, enabling alternative payment routing synchronization mechanism (PTLC), replacing vulnerable HTLC; Taproot, enabling stronger privacy properties and more compact witnesses; client-side-validation applications, including RGB smart contracts; different complex Bitcoin finance applications (BiFi), like descrete log contracts, lightspeed protocol and many others. The current standard is aiming at generalizing concept of lightning payment channels as a part of LNP (lightning network protocol) family of standards, leveraging new set of functionality offered by Bifrost protocol.

Design

First, we separate the core of the payment channel operations from other channel-related protocols, like payment routing. This enables us real-time upgradability and composability for channel structure and use of different protocols within the scope of the same channel (like HTLC, PTLC etc).

Second, we update channel transactions to use Taproot, Schnorr signatures, MuSig, TapScript – and enable sign-to-contract and pay-to-contract commitments required for client-side-validation (like RGB smart contracts).

Third, we genralize the structure of payment channel transactions for an arbitrary number of participants.

Finally, we re-define a set of messages controlling channel creation and oprations using channel proposals, strict encoding and other advancements brought by un underlying Bifrost protocol (LNPBP-50 and 51).

Specification

Funding transaction

Transaction containing multiple inputs, each of which MUST signal S2C commitment type.

Output descriptor: tr(musig(KEY_LOCAL_FUNDING, KEY_REMOTE_1_FUNDING, KEY_REMOTE_2_FUDING, ...))

Commitment transaction

Input:

Spends funding output
nSeq: 0x80800000 + COMMITMENT_NO & 0xFFFF (signals S2C DBC)
Witness stack:
- KEY_LOCAL_FUNDING + KEY_REMOTE_1_FUNDING + KEY_REMOTE_2_FUNDING + ...
- MUSIG*(KEY_LOCAL_FUNDING + KEY_REMOTE_1_FUNDING + KEY_REMOTE_2_FUNDING + ...) where * denotes optional presence of sign-to-contract commitment

Outputs:

to_local: tr(KEY_REVOCATION, and_v(v:pk(KEY_LOCAL),older(SELF_DELAY)))
- KEY_LOCAL: proposed by the local node
- SELF_DELAY: proposed by the remote node, may be rejected by local
to_remote_n: tr(KEY_UNSPENDABLE, and_v(v:pk(KEY_REMOTE_N),older(REMOTE_N_DELAY))
- KEY_REMOTE_N: propsed by the remote node N
- KEY_UNSPENDABLE: a deterministically-unspendable key computed as KEY_REMOTE_N + SHA256T("lnpbp53:unspendable", KEY_REMOTE_N) * G
- REMOTE_N_DELAY: proposed by the channel coordinator, may be rejected by remote node N
anchor: tr(KEY_UNSPENDABLE, {v:pk(KEY_LOCAL), v:pk(KEY_REMOTE_N)})
- The ooutput must contain a fixed amount of satoshis above dust limit proposed by the node adding this output to the channel proposal

Messaging

TBD

Compatibility

Upgrade

Any legacy lightning network channel can be upgraded into Bifrost channel using UPGRADE Bifrost message.

Downgrade

Channel may be downgraded from Bifrost to legacy Lightning network if and only if

Channel is a 2-peer channel
It contains only HTLC extensions and no other types of extensions
It is not used in creation of other channels
It originates from an on-chain funding transaction
It does not contain RGB assets or other protocols requiring use of deterministic bitcoin commitments or single-use-seals

Bifrost channel breaking one of these requirements may be made downgradable by removing channel protocols until all of the the requirements are met.

Channel resulting from downgrade will always use anchor_output BOLT-9 feature flag.

Rationale

Funding transaction inputs use sign-to-contract scheme

With pay-to-contract commitments if any of channel participants have the same RGB asset or other P2C-committed protocol data they will be required to merge their state transitions, expose the private asset data for all other participants and exchange full RGB consignments before signing funding transaction. This creates significant privacy leak, more attach vectors and requires much more complex protocol design.

Channel downgrade

Bifrost is a highly experimental protocol. Ability to downgrade channel to legacy stable lightning network provides a failback mechanism for cases when an attack on Bifrost channels was discovered and participating parties do not want to close the channel.