This week’s newsletter summarizes a new offchain payment resolution protocol and links to papers about potential IP-layer tracking and censorship of LN payments. Also included are announcements of new releases and release candidates (including security critical updates for BTCPay Server) and descriptions of notable changes to popular Bitcoin infrastructure software.

News

  • ● MAD-based offchain payment resolution (OPR) protocol: John Law posted to Delving Bitcoin the description of a micropayment protocol that requires both participants to contribute funds to a bond that can be effectively destroyed at any time by either participant. This creates an incentive for both parties to appease the other or risk mutually assured destruction (MAD) of the bonded funds.

    This differs from the ideal of a trustless protocol where only the party at fault loses funds in the case of a protocol breach. However, in practice, deployed trustless protocols like LN often require the conforming party to pay onchain transaction fees to recover their funds from a breach. Law uses this fact to argue for some of the benefits of a MAD-based protocol:

    • In the case of destruction of funds, it uses much less blockchain space than trustless contract enforcement, improving scalability.

    • Being based on counterparty appeasement rather than global consensus, the MAD-based protocol can enforce expiries as short as a fraction of a second rather than a minimum of at least several blocks. Law gives the example of guaranteed payment resolution (success or failure) in less than ten seconds compared to LN currently needing up to two weeks in the worst case to settle a payment.

    • In the case of a prolonged communication failure between two counterparties, the MAD-based protocol doesn’t require either party to put any data onchain (and both parties are incentivized not to do so, as they will lose their portion of the bond deposit). In a protocol like LN-Penalty, pending HTLCs in a channel must be settled onchain by a deadline if communication breaks down.

      Law emphasizes that this can make OPR much more efficient inside a channel factory, timeout tree, or other nested structure that would ideally keep the nested portions offchain.

    Matt Morehouse replied that appeasement can logically lead to slow theft. For example, Mallory claims Bob failed at an operation that’s worth 5% of the bond; Bob isn’t sure he failed, but paying Mallory 5% is better than losing his 50% of the bond, so he acquiesces; Mallory repeats. This problem is made worse by the inability to prove fault in typical communication networks: if Mallory and Bob lose contact for long enough for a failure to occur, they may each blame the other, resulting in MAD. Morehouse additionally notes that OPR requires reserving more user funds for the bond deposits, which may degrade UX—users today are already confused about the BOLT2 channel reserve that prevents them from spending more than 99% of channel balance.

    Discussion was ongoing at the time of writing.

  • ● Papers about IP-layer censorship of LN payments: Charmaine Ndolo posted to Delving Bitcoin summaries of two recent papers about reducing the privacy of LN payments and potentially censoring them. The papers note that metadata about TCP/IP packets containing LN protocol messages, such as the number of packets and the total amount of data, makes it relatively easy to guess what type of payload those messages contain (e.g. a new HTLC). An attacker who controls a network used by several nodes may be able to observe the message hopping between nodes. If the attacker also controls one of those LN nodes, it will learn some information about the message being transmitted (e.g. payment amount or that it’s an onion message). This can be used to selectively prevent some payments from succeeding, or even from failing quickly—preventing immediate retrying and potentially forcing channels to be closed onchain.

    No replies have been posted as of this writing.

Releases and release candidates

New releases and release candidates for popular Bitcoin infrastructure projects. Please consider upgrading to new releases or helping to test release candidates.

  • ● BTCPay Server 2.0.3 and 1.13.7 are maintenance releases that include security critical fixes for users of certain plugins and features. Please see the linked release notes for details.

Notable code and documentation changes

Notable recent changes in Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), Lightning BOLTs, Lightning BLIPs, Bitcoin Inquisition, and BINANAs.

  • ● Bitcoin Core #30592 removes the mempoolfullrbf setting option that allows users to disable full RBF and revert to opt-in RBF. Now that full RBF is widely adopted, there’s no benefit to disabling it, so the option has been removed. Full RBF was recently enabled by default (see Newsletter #315).

  • ● Bitcoin Core #30930 adds a peer services column to the netinfo command and an outonly filter option to display only outgoing connections. The new peer services column lists the services supported by each peer, including full blockchain data (n), bloom filters (b), segwit (w), compact filters (c), limited blockchain data up to the last 288 blocks (l), version 2 p2p transport protocol (2). Some help text updates are also made.

  • ● LDK #3283 implements BIP353 by adding support for payments to DNS-based human-readable Bitcoin payment instructions that resolve to BOLT12 offers as specified in BLIP32. A new pay_for_offer_from_human_readable_name method is added to ChannelManager to allow users to initiate payments directly to HRNs. The PR also introduces an AwaitingOffer payment state to handle pending resolutions, and a new lightning-dns-resolver crate to handle BLIP32 queries. See Newsletter #324 for previous work on this.

  • ● LND #7762 updates several lncli RPC commands to respond with status messages instead of returning empty responses, to more clearly indicate that the command was successfully executed. The affected commands include wallet releaseoutput, wallet accounts import-pubkey, wallet labeltx, sendcustom, connect, disconnect, stop, deletepayments, abandonchannel, restorechanbackup, and verifychanbackup.

Want more?

For more discussion about the topics mentioned in this newsletter, join us for the weekly Bitcoin Optech Recap on Twitter Spaces at 15:30 UTC on November 19. The discussion is also recorded and will be available from our podcasts page.