LN was designed before Schnorr signatures and multi party signature protocols like MuSig were available on bitcoin.
How can we design LN with these new tools available to us?
What if we can do away with HTLCs completely by using signature aggregation? If a protocol was possible what would look like and what will be advantages and disadvantages of such a protocol over the current HTLC based approach?
This document presents one possible construction that uses MuSig2 between multiple channels across parties to make reliable atomic payments.
- Payments should be atomic - either all parties update channel state, or none do.
- Payments should be reliable - either the sender receives a success or a failure. That is, a sender is never left hanging about the status of the payment.
- Payments should be synchronous - payment processing should complete in under seven seconds for the sender, or it should fail.
- Payment should not leak information - i.e. payments should help retain sender and receiver privacy - even from parties that forward the payment.
Continue reading the specs here
Github Discussions for any comments and feedback.
Input multi channel outputs into a single commitment that can be spent using MuSig2 signature. See figure below.
A "broadcast" built using unicast over onion routing. Turns out this is not so expensive. See spec document for details.
