This week’s newsletter summarizes a proposal to allow casual LN users to stay offline for up to several months at a time and describes a document about allowing transaction information servers to host unused wallet addresses. Also included are our regular sections with the summary of a Bitcoin Core PR Review Club, announcements of new software releases and release candidates (including a critical LND fix), and descriptions of notable changes to popular Bitcoin infrastructure software.
● LN with long timeouts proposal: John Law posted to the Lightning-Dev mailing list a proposal to allow casual Lightning users to remain offline for up to several months without risk of losing any funds to their channel partners. Although this is technically possible in the current LN protocol, it would depend on setting settlement-delay parameters to high values that would allow a griefing user or an accident to prevent funds in more than a dozen channels from being used for those same months. Law’s proposal mitigates that problem through two protocol modifications:
● Triggered HTLCs: a standard HTLC used for payment has Alice offering Bob some amount of BTC if he’s able to publish a previously-unknown preimage for a known hash digest. Alternatively, if Bob doesn’t publish the preimage by a certain time, Alice is able to spend the money back to her own wallet.
Law suggests that Bob still be allowed to claim the payment at any moment with the publication of the preimage, but Alice would need to fulfill an additional restriction. She would need to clearly warn Bob of her intent to spend the money back to her wallet by getting a trigger transaction confirmed onchain. Only when the trigger transaction had been confirmed by a certain number of blocks (or for a certain duration of time) would Alice be able to spend the money.
This would ensure Bob was able to claim his funds at any time up until the trigger transaction had received the agreed-upon number of confirmations, even if months had passed since a normal HTLC would’ve timed out. If Bob is adequately compensated for his waiting, then it’s ok if Alice remains offline all that time. For an HTLC routed from Alice through Bob onto some distant node, only the channel between Alice and Bob would be affected—every other channel would settle the HTLC promptly (as in the current LN protocol).
● Asymmetric delayed commitment transactions: each of the two partners in an LN channel holds an unpublished commitment transaction that they can publish and try to get confirmed at any time. Both versions of the transaction spend the same UTXO, so they conflict with each other—meaning only one can actually get confirmed.
This means when Alice wants to close the channel, she can’t just simply broadcast her version of the commitment transaction with a reasonable feerate and assume it will get confirmed. She also has to wait and check whether Bob instead gets his version of the commitment transaction confirmed, in which case she may need to take additional actions to verify his transaction included the latest channel state.
Law proposes that Alice’s version of the commitment transaction remain the same as today so that she can publish it at any time, but that Bob’s version include a time lock so that he can only publish it if Alice has been inactive for a long time. Ideally, this allows Alice to publish the latest state secure in the knowledge that Bob can’t publish a contradictory version, allowing her to safely go offline after her publication.
● Recommendations for unique address servers: Ruben Somsen posted to the Bitcoin-Dev mailing list a document with another suggestion for how users can avoid output linking without trusting a third-party service or using an cryptographic protocol that’s not currently widely supported, like BIP47 or silent payments. The recommended method is especially intended for wallets that are already providing their addresses to third parties, such as those that use public address lookup servers (which is believed to be the majority of lightweight wallets).
For an example of how the method might work, Alice’s wallet registers 100 addresses on the Example.com electrum-style server. She then includes “example.com/alice” in her email signature. When Bob wants to donate money to Alice, he visits her URL, gets an address, verifies that Alice signed it, and then pays to it.
The idea has the advantage of being widely compatible with many wallets through a partly-manual process and possibly easy to implement with an automated process. Its downside is that users who are already compromising their privacy by sharing addresses with a server will be further committing to the privacy loss.
Bitcoin Core PR Review Club
In this monthly section, we summarize a recent Bitcoin Core PR Review Club meeting, highlighting some of the important questions and answers. Click on a question below to see a summary of the answer from the meeting.
Make AddrFetch connections to fixed seeds
is a PR by Martin Zumsande that makes
AddrFetch connections to
the fixed seeds (hard-coded IP addresses) instead of just adding
AddrMan (the database of our peers).
When a new node starts up from scratch, it must first connect with some peers from whom it will perform Initial Block Download (IBD). Under what circumstances does it connect to the fixed seeds?
Only if it isn’t able to connect to peers whose addresses are provided by the hard-coded Bitcoin DNS seed nodes. This most commonly occurs when the node is configured to not use IPv4 or IPv6 (for example,
What observable behavior change does this PR introduce? What kinds of addresses do we add to
AddrMan, and under what circumstances?
The node, rather than immediately adding the fixed seeds to its
AddrManand making full connections to some of them, instead makes
AddrFetchconnections to some of them, and adds the returned addresses to
AddrFetchare short-term connections that are used only to fetch addresses.) The node then connects to some of the addresses now in its
AddrManto perform IBD. This results in fewer full connections to the fixed-seed nodes; instead, more connections are attempted from the much larger set of nodes that the fixed-seed nodes tell us about.
AddrFetchconnections can return any type of addresses, for example,
tor; the results are not limited to IPv4 and IPv6. ➚
Why might we want to make an
AddrFetchconnection instead of a full outbound connection to fixed seeds? Why might the node operator behind a fixed seed prefer this as well?
AddrFetchconnection allows our node to choose IBD peers from a much larger set of peers, which increases overall network connectivity distribution. The fixed seed node operators would be less likely to have multiple simultaneous IBD peers, which reduces the resource requirements on their nodes. ➚
The DNS seed nodes are expected to be responsive and serve up-to-date addresses of Bitcoin nodes. Why doesn’t this help a
The DNS seed nodes provide only IPv4 and IPv6 addresses; they’re not able to supply any other type of address. ➚
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.
Notable code and documentation changes
Notable changes this week in Bitcoin Core, Core Lightning, Eclair, LDK, LND, libsecp256k1, Hardware Wallet Interface (HWI), Rust Bitcoin, BTCPay Server, BDK, Bitcoin Improvement Proposals (BIPs), and Lightning BOLTs.
- ● LND #6500 Adds the ability to encrypt the Tor private key on disk
using the wallet’s private key instead of storing it in plaintext.
Using the flag,
--tor.encryptkey, LND encrypts the private key and the encrypted blob is written to the same file on disk, allowing users to still keep the same functionality (like refreshing a hidden service), but adds protection when running in untrusted environments.