mirrors/lightning-bolts
1
0
Fork 0
mirror of https://github.com/lightning/bolts.git synced 2024-11-19 01:50:03 +01:00
Code Issues Projects Releases Wiki Activity
e7017173d6
lightning-bolts/02-peer-protocol.md
Vincenzo Palazzo e7017173d6
bolt2: disallow sending multiple shutdown msg (#977) 
The rationale for this is to avoid bad cases like the following one
which was previously allowed:

* sender -> shutdown(script_one) -> receiver
* sender -> shutdown(script_two) -> receiver
* sender <- shutdown(script_one) <- receiver

Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
2022-05-18 15:51:37 +02:00

74 KiB
Raw Blame History

BOLT #2: Peer Protocol for Channel Management

The peer channel protocol has three phases: establishment, normal operation, and closing.

Table of Contents

Channel

Definition of channel_id

Some messages use a channel_id to identify the channel. It's derived from the funding transaction by combining the funding_txid and the funding_output_index, using big-endian exclusive-OR (i.e. funding_output_index alters the last 2 bytes).

Prior to channel establishment, a temporary_channel_id is used, which is a random nonce.

Note that as duplicate temporary_channel_ids may exist from different peers, APIs which reference channels by their channel id before the funding transaction is created are inherently unsafe. The only protocol-provided identifier for a channel before funding_created has been exchanged is the (source_node_id, destination_node_id, temporary_channel_id) tuple. Note that any such APIs which reference channels by their channel id before the funding transaction is confirmed are also not persistent - until you know the script pubkey corresponding to the funding output nothing prevents duplicative channel ids.

Channel Establishment

After authenticating and initializing a connection (BOLT #8 and BOLT #1, respectively), channel establishment may begin. This consists of the funding node (funder) sending an open_channel message, followed by the responding node (fundee) sending accept_channel. With the channel parameters locked in, the funder is able to create the funding transaction and both versions of the commitment transaction, as described in BOLT #3. The funder then sends the outpoint of the funding output with the funding_created message, along with the signature for the fundee's version of the commitment transaction. Once the fundee learns the funding outpoint, it's able to generate the signature for the funder's version of the commitment transaction and send it over using the funding_signed message.

Once the channel funder receives the funding_signed message, it must broadcast the funding transaction to the Bitcoin network. After the funding_signed message is sent/received, both sides should wait for the funding transaction to enter the blockchain and reach the specified depth (number of confirmations). After both sides have sent the funding_locked message, the channel is established and can begin normal operation. The funding_locked message includes information that will be used to construct channel authentication proofs.

    +-------+                              +-------+
    |       |--(1)---  open_channel  ----->|       |
    |       |<-(2)--  accept_channel  -----|       |
    |       |                              |       |
    |   A   |--(3)--  funding_created  --->|   B   |
    |       |<-(4)--  funding_signed  -----|       |
    |       |                              |       |
    |       |--(5)--- funding_locked  ---->|       |
    |       |<-(6)--- funding_locked  -----|       |
    +-------+                              +-------+

    - where node A is 'funder' and node B is 'fundee'

If this fails at any stage, or if one node decides the channel terms offered by the other node are not suitable, the channel establishment fails.

Note that multiple channels can operate in parallel, as all channel messages are identified by either a temporary_channel_id (before the funding transaction is created) or a channel_id (derived from the funding transaction).

The open_channel Message

This message contains information about a node and indicates its desire to set up a new channel. This is the first step toward creating the funding transaction and both versions of the commitment transaction.

  1. type: 32 (open_channel)

  2. data:

    • [chain_hash:chain_hash]
    • [32*byte:temporary_channel_id]
    • [u64:funding_satoshis]
    • [u64:push_msat]
    • [u64:dust_limit_satoshis]
    • [u64:max_htlc_value_in_flight_msat]
    • [u64:channel_reserve_satoshis]
    • [u64:htlc_minimum_msat]
    • [u32:feerate_per_kw]
    • [u16:to_self_delay]
    • [u16:max_accepted_htlcs]
    • [point:funding_pubkey]
    • [point:revocation_basepoint]
    • [point:payment_basepoint]
    • [point:delayed_payment_basepoint]
    • [point:htlc_basepoint]
    • [point:first_per_commitment_point]
    • [byte:channel_flags]
    • [open_channel_tlvs:tlvs]

  3. tlv_stream: open_channel_tlvs

  4. types:

    1. type: 0 (upfront_shutdown_script)
    2. data:
      • [...*byte:shutdown_scriptpubkey]
    3. type: 1 (channel_type)
    4. data:
      • [...*byte:type]

The chain_hash value denotes the exact blockchain that the opened channel will reside within. This is usually the genesis hash of the respective blockchain. The existence of the chain_hash allows nodes to open channels across many distinct blockchains as well as have channels within multiple blockchains opened to the same peer (if it supports the target chains).

The temporary_channel_id is used to identify this channel on a per-peer basis until the funding transaction is established, at which point it is replaced by the channel_id, which is derived from the funding transaction.

funding_satoshis is the amount the sender is putting into the channel. push_msat is an amount of initial funds that the sender is unconditionally giving to the receiver. dust_limit_satoshis is the threshold below which outputs should not be generated for this node's commitment or HTLC transactions (i.e. HTLCs below this amount plus HTLC transaction fees are not enforceable on-chain). This reflects the reality that tiny outputs are not considered standard transactions and will not propagate through the Bitcoin network. channel_reserve_satoshis is the minimum amount that the other node is to keep as a direct payment. htlc_minimum_msat indicates the smallest value HTLC this node will accept.

max_htlc_value_in_flight_msat is a cap on total value of outstanding HTLCs, which allows a node to limit its exposure to HTLCs; similarly, max_accepted_htlcs limits the number of outstanding HTLCs the other node can offer.

feerate_per_kw indicates the initial fee rate in satoshi per 1000-weight (i.e. 1/4 the more normally-used 'satoshi per 1000 vbytes') that this side will pay for commitment and HTLC transactions, as described in BOLT #3 (this can be adjusted later with an update_fee message).

to_self_delay is the number of blocks that the other node's to-self outputs must be delayed, using OP_CHECKSEQUENCEVERIFY delays; this is how long it will have to wait in case of breakdown before redeeming its own funds.

funding_pubkey is the public key in the 2-of-2 multisig script of the funding transaction output.

The various _basepoint fields are used to derive unique keys as described in BOLT #3 for each commitment transaction. Varying these keys ensures that the transaction ID of each commitment transaction is unpredictable to an external observer, even if one commitment transaction is seen; this property is very useful for preserving privacy when outsourcing penalty transactions to third parties.

first_per_commitment_point is the per-commitment point to be used for the first commitment transaction,

Only the least-significant bit of channel_flags is currently defined: announce_channel. This indicates whether the initiator of the funding flow wishes to advertise this channel publicly to the network, as detailed within BOLT #7.

The shutdown_scriptpubkey allows the sending node to commit to where funds will go on mutual close, which the remote node should enforce even if a node is compromised later.

The option_support_large_channel is a feature used to let everyone know this node will accept funding_satoshis greater than or equal to 2^24. Since it's broadcast in the node_announcement message other nodes can use it to identify peers willing to accept large channel even before exchanging the init message with them.

Defined Channel Types

Channel types are an explicit enumeration: for convenience of future definitions they reuse even feature bits, but they are not an arbitrary combination (they represent the persistent features which affect the channel operation).

The currently defined types are:

  • no features (no bits set)
  • option_static_remotekey (bit 12)
  • option_anchor_outputs and option_static_remotekey (bits 20 and 12)
  • option_anchors_zero_fee_htlc_tx and option_static_remotekey (bits 22 and 12)

Requirements

The sending node:

  • MUST ensure the chain_hash value identifies the chain it wishes to open the channel within.
  • MUST ensure temporary_channel_id is unique from any other channel ID with the same peer.
  • if both nodes advertised option_support_large_channel:
    • MAY set funding_satoshis greater than or equal to 2^24 satoshi.
  • otherwise:
    • MUST set funding_satoshis to less than 2^24 satoshi.
  • MUST set push_msat to equal or less than 1000 * funding_satoshis.
  • MUST set funding_pubkey, revocation_basepoint, htlc_basepoint, payment_basepoint, and delayed_payment_basepoint to valid secp256k1 pubkeys in compressed format.
  • MUST set first_per_commitment_point to the per-commitment point to be used for the initial commitment transaction, derived as specified in BOLT #3.
  • MUST set channel_reserve_satoshis greater than or equal to dust_limit_satoshis.
  • MUST set undefined bits in channel_flags to 0.
  • if both nodes advertised the option_upfront_shutdown_script feature:
    • MUST include upfront_shutdown_script with either a valid shutdown_scriptpubkey as required by shutdown scriptpubkey, or a zero-length shutdown_scriptpubkey (ie. 0x0000).
  • otherwise:
    • MAY include upfront_shutdown_script.
  • if it includes open_channel_tlvs:
    • MUST include upfront_shutdown_script.
  • if option_channel_type is negotiated:
    • MUST set channel_type
  • if it includes channel_type:
    • MUST set it to a defined type representing the type it wants.
    • MUST use the smallest bitmap possible to represent the channel type.
    • SHOULD NOT set it to a type containing a feature which was not negotiated.

The sending node SHOULD:

  • set to_self_delay sufficient to ensure the sender can irreversibly spend a commitment transaction output, in case of misbehavior by the receiver.
  • set feerate_per_kw to at least the rate it estimates would cause the transaction to be immediately included in a block.
  • set dust_limit_satoshis to a sufficient value to allow commitment transactions to propagate through the Bitcoin network.
  • set htlc_minimum_msat to the minimum value HTLC it's willing to accept from this peer.

The receiving node MUST:

  • ignore undefined bits in channel_flags.
  • if the connection has been re-established after receiving a previous open_channel, BUT before receiving a funding_created message:
    • accept a new open_channel message.
    • discard the previous open_channel message.

The receiving node MAY fail the channel if:

  • option_channel_type was negotiated but the message doesn't include a channel_type
  • announce_channel is false (0), yet it wishes to publicly announce the channel.
  • funding_satoshis is too small.
  • it considers htlc_minimum_msat too large.
  • it considers max_htlc_value_in_flight_msat too small.
  • it considers channel_reserve_satoshis too large.
  • it considers max_accepted_htlcs too small.
  • it considers dust_limit_satoshis too large.

The receiving node MUST fail the channel if:

  • the chain_hash value is set to a hash of a chain that is unknown to the receiver.
  • push_msat is greater than funding_satoshis * 1000.
  • to_self_delay is unreasonably large.
  • max_accepted_htlcs is greater than 483.
  • it considers feerate_per_kw too small for timely processing or unreasonably large.
  • funding_pubkey, revocation_basepoint, htlc_basepoint, payment_basepoint, or delayed_payment_basepoint are not valid secp256k1 pubkeys in compressed format.
  • dust_limit_satoshis is greater than channel_reserve_satoshis.
  • dust_limit_satoshis is smaller than 354 satoshis (see BOLT 3).
  • the funder's amount for the initial commitment transaction is not sufficient for full fee payment.
  • both to_local and to_remote amounts for the initial commitment transaction are less than or equal to channel_reserve_satoshis (see BOLT 3).
  • funding_satoshis is greater than or equal to 2^24 and the receiver does not support option_support_large_channel.
  • It supports channel_type, channel_type was set, and the type is not suitable.

The receiving node MUST NOT:

  • consider funds received, using push_msat, to be received until the funding transaction has reached sufficient depth.

Rationale

The requirement for funding_satoshis to be less than 2^24 satoshi was a temporary self-imposed limit while implementations were not yet considered stable, it can be lifted by advertising option_support_large_channel.

The channel reserve is specified by the peer's channel_reserve_satoshis: 1% of the channel total is suggested. Each side of a channel maintains this reserve so it always has something to lose if it were to try to broadcast an old, revoked commitment transaction. Initially, this reserve may not be met, as only one side has funds; but the protocol ensures that there is always progress toward meeting this reserve, and once met, it is maintained.

The sender can unconditionally give initial funds to the receiver using a non-zero push_msat, but even in this case we ensure that the funder has sufficient remaining funds to pay fees and that one side has some amount it can spend (which also implies there is at least one non-dust output). Note that, like any other on-chain transaction, this payment is not certain until the funding transaction has been confirmed sufficiently (with a danger of double-spend until this occurs) and may require a separate method to prove payment via on-chain confirmation.

The feerate_per_kw is generally only of concern to the sender (who pays the fees), but there is also the fee rate paid by HTLC transactions; thus, unreasonably large fee rates can also penalize the recipient.

Separating the htlc_basepoint from the payment_basepoint improves security: a node needs the secret associated with the htlc_basepoint to produce HTLC signatures for the protocol, but the secret for the payment_basepoint can be in cold storage.

The requirement that channel_reserve_satoshis is not considered dust according to dust_limit_satoshis eliminates cases where all outputs would be eliminated as dust. The similar requirements in accept_channel ensure that both sides' channel_reserve_satoshis are above both dust_limit_satoshis.

The receiver should not accept large dust_limit_satoshis, as this could be used in griefing attacks, where the peer publishes its commitment with a lot of dust htlcs, which effectively become miner fees.

Details for how to handle a channel failure can be found in BOLT 5:Failing a Channel.

The accept_channel Message

This message contains information about a node and indicates its acceptance of the new channel. This is the second step toward creating the funding transaction and both versions of the commitment transaction.

  1. type: 33 (accept_channel)

  2. data:

    • [32*byte:temporary_channel_id]
    • [u64:dust_limit_satoshis]
    • [u64:max_htlc_value_in_flight_msat]
    • [u64:channel_reserve_satoshis]
    • [u64:htlc_minimum_msat]
    • [u32:minimum_depth]
    • [u16:to_self_delay]
    • [u16:max_accepted_htlcs]
    • [point:funding_pubkey]
    • [point:revocation_basepoint]
    • [point:payment_basepoint]
    • [point:delayed_payment_basepoint]
    • [point:htlc_basepoint]
    • [point:first_per_commitment_point]
    • [accept_channel_tlvs:tlvs]

  3. tlv_stream: accept_channel_tlvs

  4. types:

    1. type: 0 (upfront_shutdown_script)
    2. data:
      • [...*byte:shutdown_scriptpubkey]
    3. type: 1 (channel_type)
    4. data:
      • [...*byte:type]

Requirements

The temporary_channel_id MUST be the same as the temporary_channel_id in the open_channel message.

The sender:

  • SHOULD set minimum_depth to a number of blocks it considers reasonable to avoid double-spending of the funding transaction.
  • MUST set channel_reserve_satoshis greater than or equal to dust_limit_satoshis from the open_channel message.
  • MUST set dust_limit_satoshis less than or equal to channel_reserve_satoshis from the open_channel message.
  • if option_channel_type was negotiated:
    • MUST set channel_type to the channel_type from open_channel

The receiver:

  • if minimum_depth is unreasonably large:
    • MAY reject the channel.
  • if channel_reserve_satoshis is less than dust_limit_satoshis within the open_channel message:
    • MUST reject the channel.
  • if channel_reserve_satoshis from the open_channel message is less than dust_limit_satoshis:
    • MUST reject the channel.
  • if channel_type is set, and channel_type was set in open_channel, and they are not equal types:
    • MUST reject the channel.
  • if option_channel_type was negotiated but the message doesn't include a channel_type:
    • MAY reject the channel.

Other fields have the same requirements as their counterparts in open_channel.

The funding_created Message

This message describes the outpoint which the funder has created for the initial commitment transactions. After receiving the peer's signature, via funding_signed, it will broadcast the funding transaction.

  1. type: 34 (funding_created)
  2. data:
    • [32*byte:temporary_channel_id]
    • [sha256:funding_txid]
    • [u16:funding_output_index]
    • [signature:signature]

Requirements

The sender MUST set:

  • temporary_channel_id the same as the temporary_channel_id in the open_channel message.
  • funding_txid to the transaction ID of a non-malleable transaction,
    • and MUST NOT broadcast this transaction.
  • funding_output_index to the output number of that transaction that corresponds the funding transaction output, as defined in BOLT #3.
  • signature to the valid signature using its funding_pubkey for the initial commitment transaction, as defined in BOLT #3.

The sender:

  • when creating the funding transaction:
    • SHOULD use only BIP141 (Segregated Witness) inputs.
    • SHOULD ensure the funding transaction confirms in the next 2016 blocks.

The recipient:

  • if signature is incorrect OR non-compliant with LOW-S-standard ruleLOWS:
    • MUST send a warning and close the connection, or send an error and fail the channel.

Rationale

The funding_output_index can only be 2 bytes, since that's how it's packed into the channel_id and used throughout the gossip protocol. The limit of 65535 outputs should not be overly burdensome.

A transaction with all Segregated Witness inputs is not malleable, hence the funding transaction recommendation.

The funder may use CPFP on a change output to ensure that the funding transaction confirms before 2016 blocks, otherwise the fundee may forget that channel.

The funding_signed Message

This message gives the funder the signature it needs for the first commitment transaction, so it can broadcast the transaction knowing that funds can be redeemed, if need be.

This message introduces the channel_id to identify the channel. It's derived from the funding transaction by combining the funding_txid and the funding_output_index, using big-endian exclusive-OR (i.e. funding_output_index alters the last 2 bytes).

  1. type: 35 (funding_signed)
  2. data:
    • [channel_id:channel_id]
    • [signature:signature]

Requirements

Both peers:

  • if channel_type was present in both open_channel and accept_channel:
    • This is the channel_type (they must be equal, required above)
  • otherwise:
    • if option_anchors_zero_fee_htlc_tx was negotiated:
      • the channel_type is option_anchors_zero_fee_htlc_tx and option_static_remotekey (bits 22 and 12)
    • otherwise, if option_anchor_outputs was negotiated:
      • the channel_type is option_anchor_outputs and option_static_remotekey (bits 20 and 12)
    • otherwise, if option_static_remotekey was negotiated:
      • the channel_type is option_static_remotekey (bit 12)
    • otherwise:
      • the channel_type is empty
  • MUST use that channel_type for all commitment transactions.

The sender MUST set:

  • channel_id by exclusive-OR of the funding_txid and the funding_output_index from the funding_created message.
  • signature to the valid signature, using its funding_pubkey for the initial commitment transaction, as defined in BOLT #3.

The recipient:

  • if signature is incorrect OR non-compliant with LOW-S-standard ruleLOWS:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • MUST NOT broadcast the funding transaction before receipt of a valid funding_signed.
  • on receipt of a valid funding_signed:
    • SHOULD broadcast the funding transaction.

Rationale

We decide on option_static_remotekey, option_anchor_outputs or option_anchors_zero_fee_htlc_tx at this point when we first have to generate the commitment transaction. The feature bits that were communicated in the init message exchange for the current connection determine the channel commitment format for the total lifetime of the channel. Even if a later reconnection does not negotiate this parameter, this channel will continue to use option_static_remotekey, option_anchor_outputs or option_anchors_zero_fee_htlc_tx; we don't support "downgrading".

option_anchors_zero_fee_htlc_tx is considered superior to option_anchor_outputs, which again is considered superior to option_static_remotekey, and the superior one is favored if more than one is negotiated.

The funding_locked Message

This message indicates that the funding transaction has reached the minimum_depth asked for in accept_channel. Once both nodes have sent this, the channel enters normal operating mode.

  1. type: 36 (funding_locked)
  2. data:
    • [channel_id:channel_id]
    • [point:next_per_commitment_point]

Requirements

The sender MUST:

  • NOT send funding_locked unless outpoint of given by funding_txid and funding_output_index in the funding_created message pays exactly funding_satoshis to the scriptpubkey specified in BOLT #3.
  • wait until the funding transaction has reached minimum_depth before sending this message.
  • set next_per_commitment_point to the per-commitment point to be used for the following commitment transaction, derived as specified in BOLT #3.

A non-funding node (fundee):

  • SHOULD forget the channel if it does not see the correct funding transaction after a timeout of 2016 blocks.

From the point of waiting for funding_locked onward, either node MAY send an error and fail the channel if it does not receive a required response from the other node after a reasonable timeout.

Rationale

The non-funder can simply forget the channel ever existed, since no funds are at risk. If the fundee were to remember the channel forever, this would create a Denial of Service risk; therefore, forgetting it is recommended (even if the promise of push_msat is significant).

If the fundee forgets the channel before it was confirmed, the funder will need to broadcast the commitment transaction to get his funds back and open a new channel. To avoid this, the funder should ensure the funding transaction confirms in the next 2016 blocks.

Channel Close

Nodes can negotiate a mutual close of the connection, which unlike a unilateral close, allows them to access their funds immediately and can be negotiated with lower fees.

Closing happens in two stages:

  1. one side indicates it wants to clear the channel (and thus will accept no new HTLCs)

  2. once all HTLCs are resolved, the final channel close negotiation begins.

     +-------+                              +-------+
 |       |--(1)-----  shutdown  ------->|       |
 |       |<-(2)-----  shutdown  --------|       |
 |       |                              |       |
 |       | <complete all pending HTLCs> |       |
 |   A   |                 ...          |   B   |
 |       |                              |       |
 |       |--(3)-- closing_signed  F1--->|       |
 |       |<-(4)-- closing_signed  F2----|       |
 |       |              ...             |       |
 |       |--(?)-- closing_signed  Fn--->|       |
 |       |<-(?)-- closing_signed  Fn----|       |
 +-------+                              +-------+

Closing Initiation: shutdown

Either node (or both) can send a shutdown message to initiate closing, along with the scriptpubkey it wants to be paid to.

  1. type: 38 (shutdown)
  2. data:
    • [channel_id:channel_id]
    • [u16:len]
    • [len*byte:scriptpubkey]

Requirements

A sending node:

  • if it hasn't sent a funding_created (if it is a funder) or a funding_signed (if it is a fundee):

    • MUST NOT send a shutdown

  • MAY send a shutdown before a funding_locked, i.e. before the funding transaction has reached minimum_depth.

  • if there are updates pending on the receiving node's commitment transaction:

    • MUST NOT send a shutdown.

  • MUST NOT send multiple shutdown messages.

  • MUST NOT send an update_add_htlc after a shutdown.

  • if no HTLCs remain in either commitment transaction:

    • MUST NOT send any update message after a shutdown.

  • SHOULD fail to route any HTLC added after it has sent shutdown.

  • if it sent a non-zero-length shutdown_scriptpubkey in open_channel or accept_channel:

    • MUST send the same value in scriptpubkey.

  • MUST set scriptpubkey in one of the following forms:

    1. OP_0 20 20-bytes (version 0 pay to witness pubkey hash), OR
    2. OP_0 32 32-bytes (version 0 pay to witness script hash), OR
    3. if (and only if) option_shutdown_anysegwit is negotiated:
    • OP_1 through OP_16 inclusive, followed by a single push of 2 to 40 bytes (witness program versions 1 through 16)

A receiving node:

  • if it hasn't received a funding_signed (if it is a funder) or a funding_created (if it is a fundee):
    • SHOULD send an error and fail the channel.
  • if the scriptpubkey is not in one of the above forms:
    • SHOULD send a warning.
  • if it hasn't sent a funding_locked yet:
    • MAY reply to a shutdown message with a shutdown
  • once there are no outstanding updates on the peer, UNLESS it has already sent a shutdown:
    • MUST reply to a shutdown message with a shutdown
  • if both nodes advertised the option_upfront_shutdown_script feature, and the receiving node received a non-zero-length shutdown_scriptpubkey in open_channel or accept_channel, and that shutdown_scriptpubkey is not equal to scriptpubkey:
    • MAY send a warning.
    • MUST fail the connection.

Rationale

If channel state is always "clean" (no pending changes) when a shutdown starts, the question of how to behave if it wasn't is avoided: the sender always sends a commitment_signed first.

As shutdown implies a desire to terminate, it implies that no new HTLCs will be added or accepted. Once any HTLCs are cleared, the peer may immediately begin closing negotiation, so we ban further updates to the commitment transaction (in particular, update_fee would be possible otherwise).

The scriptpubkey forms include only standard segwit forms accepted by the Bitcoin network, which ensures the resulting transaction will propagate to miners. However old nodes may send non-segwit scripts, which may be accepted for backwards-compatibility (with a caveat to force-close if this output doesn't meet dust relay requirements).

The option_upfront_shutdown_script feature means that the node wanted to pre-commit to shutdown_scriptpubkey in case it was compromised somehow. This is a weak commitment (a malevolent implementation tends to ignore specifications like this one!), but it provides an incremental improvement in security by requiring the cooperation of the receiving node to change the scriptpubkey.

The shutdown response requirement implies that the node sends commitment_signed to commit any outstanding changes before replying; however, it could theoretically reconnect instead, which would simply erase all outstanding uncommitted changes.

Closing Negotiation: closing_signed

Once shutdown is complete and the channel is empty of HTLCs, the final current commitment transactions will have no HTLCs, and closing fee negotiation begins. The funder chooses a fee it thinks is fair, and signs the closing transaction with the scriptpubkey fields from the shutdown messages (along with its chosen fee) and sends the signature; the other node then replies similarly, using a fee it thinks is fair. This exchange continues until both agree on the same fee or when one side fails the channel.

In the modern method, the funder sends its permissible fee range, and the non-funder has to pick a fee in this range. If the non-funder chooses the same value, negotiation is complete after two messages, otherwise the funder will reply with the same value (completing after three messages).

  1. type: 39 (closing_signed)

  2. data:

    • [channel_id:channel_id]
    • [u64:fee_satoshis]
    • [signature:signature]
    • [closing_signed_tlvs:tlvs]

  3. tlv_stream: closing_signed_tlvs

  4. types:

    1. type: 1 (fee_range)
    2. data:
      • [u64:min_fee_satoshis]
      • [u64:max_fee_satoshis]

Requirements

The funding node:

  • after shutdown has been received, AND no HTLCs remain in either commitment transaction:
    • SHOULD send a closing_signed message.

The sending node:

  • SHOULD set the initial fee_satoshis according to its estimate of cost of inclusion in a block.
  • SHOULD set fee_range according to the minimum and maximum fees it is prepared to pay for a close transaction.
  • if it doesn't receive a closing_signed response after a reasonable amount of time:
    • MUST fail the channel
  • if it is not the funder:
    • SHOULD set max_fee_satoshis to at least the max_fee_satoshis received
    • SHOULD set min_fee_satoshis to a fairly low value
  • MUST set signature to the Bitcoin signature of the close transaction, as specified in BOLT #3.

The receiving node:

  • if the signature is not valid for either variant of closing transaction specified in BOLT #3 OR non-compliant with LOW-S-standard ruleLOWS:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if fee_satoshis is equal to its previously sent fee_satoshis:
    • SHOULD sign and broadcast the final closing transaction.
    • MAY close the connection.
  • if fee_satoshis matches its previously sent fee_range:
    • SHOULD use fee_satoshis to sign and broadcast the final closing transaction
    • SHOULD reply with a closing_signed with the same fee_satoshis value if it is different from its previously sent fee_satoshis
    • MAY close the connection.
  • if the message contains a fee_range:
    • if there is no overlap between that and its own fee_range:
      • SHOULD send a warning
      • MUST fail the channel if it doesn't receive a satisfying fee_range after a reasonable amount of time
    • otherwise:
      • if it is the funder:
        • if fee_satoshis is not in the overlap between the sent and received fee_range:
          • MUST fail the channel
        • otherwise:
          • MUST reply with the same fee_satoshis.
      • otherwise (it is not the funder):
        • if it has already sent a closing_signed:
          • if fee_satoshis is not the same as the value it sent:
            • MUST fail the channel
        • otherwise:
          • MUST propose a fee_satoshis in the overlap between received and (about-to-be) sent fee_range.
  • otherwise, if fee_satoshis is not strictly between its last-sent fee_satoshis and its previously-received fee_satoshis, UNLESS it has since reconnected:
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
  • otherwise, if the receiver agrees with the fee:
    • SHOULD reply with a closing_signed with the same fee_satoshis value.
  • otherwise:
    • MUST propose a value "strictly between" the received fee_satoshis and its previously-sent fee_satoshis.

The receiving node:

  • if one of the outputs in the closing transaction is below the dust limit for its scriptpubkey (see BOLT 3):
    • MUST fail the channel

Rationale

When fee_range is not provided, the "strictly between" requirement ensures that forward progress is made, even if only by a single satoshi at a time. To avoid keeping state and to handle the corner case, where fees have shifted between disconnection and reconnection, negotiation restarts on reconnection.

Note there is limited risk if the closing transaction is delayed, but it will be broadcast very soon; so there is usually no reason to pay a premium for rapid processing.

Note that the non-funder is not paying the fee, so there is no reason for it to have a maximum feerate. It may want a minimum feerate, however, to ensure that the transaction propagates. It can always use CPFP later to speed up confirmation if necessary, so that minimum should be low.

It may happen that the closing transaction doesn't meet bitcoin's default relay policies (e.g. when using a non-segwit shutdown script for an output below 546 satoshis, which is possible if dust_limit_satoshis is below 546 satoshis). No funds are at risk when that happens, but the channel must be force-closed as the closing transaction will likely never reach miners.

Normal Operation

Once both nodes have exchanged funding_locked (and optionally announcement_signatures), the channel can be used to make payments via Hashed Time Locked Contracts.

Changes are sent in batches: one or more update_ messages are sent before a commitment_signed message, as in the following diagram:

    +-------+                               +-------+
    |       |--(1)---- update_add_htlc ---->|       |
    |       |--(2)---- update_add_htlc ---->|       |
    |       |<-(3)---- update_add_htlc -----|       |
    |       |                               |       |
    |       |--(4)--- commitment_signed --->|       |
    |   A   |<-(5)---- revoke_and_ack ------|   B   |
    |       |                               |       |
    |       |<-(6)--- commitment_signed ----|       |
    |       |--(7)---- revoke_and_ack ----->|       |
    |       |                               |       |
    |       |--(8)--- commitment_signed --->|       |
    |       |<-(9)---- revoke_and_ack ------|       |
    +-------+                               +-------+

Counter-intuitively, these updates apply to the other node's commitment transaction; the node only adds those updates to its own commitment transaction when the remote node acknowledges it has applied them via revoke_and_ack.

Thus each update traverses through the following states:

  1. pending on the receiver
  2. in the receiver's latest commitment transaction
  3. ... and the receiver's previous commitment transaction has been revoked, and the update is pending on the sender
  4. ... and in the sender's latest commitment transaction
  5. ... and the sender's previous commitment transaction has been revoked

As the two nodes' updates are independent, the two commitment transactions may be out of sync indefinitely. This is not concerning: what matters is whether both sides have irrevocably committed to a particular update or not (the final state, above).

Forwarding HTLCs

In general, a node offers HTLCs for two reasons: to initiate a payment of its own, or to forward another node's payment. In the forwarding case, care must be taken to ensure the outgoing HTLC cannot be redeemed unless the incoming HTLC can be redeemed. The following requirements ensure this is always true.

The respective addition/removal of an HTLC is considered irrevocably committed when:

  1. The commitment transaction with/without it is committed to by both nodes, and any previous commitment transaction without/with it has been revoked, OR
  2. The commitment transaction with/without it has been irreversibly committed to the blockchain.

Requirements

A node:

  • until an incoming HTLC has been irrevocably committed:
    • MUST NOT offer the corresponding outgoing HTLC (update_add_htlc) in response to that incoming HTLC.
  • until the removal of an outgoing HTLC is irrevocably committed, OR until the outgoing on-chain HTLC output has been spent via the HTLC-timeout transaction (with sufficient depth):
    • MUST NOT fail the incoming HTLC (update_fail_htlc) that corresponds to that outgoing HTLC.
  • once the cltv_expiry of an incoming HTLC has been reached, OR if cltv_expiry minus current_height is less than cltv_expiry_delta for the corresponding outgoing HTLC:
    • MUST fail that incoming HTLC (update_fail_htlc).
  • if an incoming HTLC's cltv_expiry is unreasonably far in the future:
    • SHOULD fail that incoming HTLC (update_fail_htlc).
  • upon receiving an update_fulfill_htlc for an outgoing HTLC, OR upon discovering the payment_preimage from an on-chain HTLC spend:
    • MUST fulfill the incoming HTLC that corresponds to that outgoing HTLC.

Rationale

In general, one side of the exchange needs to be dealt with before the other. Fulfilling an HTLC is different: knowledge of the preimage is, by definition, irrevocable and the incoming HTLC should be fulfilled as soon as possible to reduce latency.

An HTLC with an unreasonably long expiry is a denial-of-service vector and therefore is not allowed. Note that the exact value of "unreasonable" is currently unclear and may depend on network topology.

cltv_expiry_delta Selection

Once an HTLC has timed out, it can either be fulfilled or timed-out; care must be taken around this transition, both for offered and received HTLCs.

Consider the following scenario, where A sends an HTLC to B, who forwards to C, who delivers the goods as soon as the payment is received.

  1. C needs to be sure that the HTLC from B cannot time out, even if B becomes unresponsive; i.e. C can fulfill the incoming HTLC on-chain before B can time it out on-chain.

  2. B needs to be sure that if C fulfills the HTLC from B, it can fulfill the incoming HTLC from A; i.e. B can get the preimage from C and fulfill the incoming HTLC on-chain before A can time it out on-chain.

The critical settings here are the cltv_expiry_delta in BOLT #7 and the related min_final_cltv_expiry in BOLT #11. cltv_expiry_delta is the minimum difference in HTLC CLTV timeouts, in the forwarding case (B). min_final_cltv_expiry is the minimum difference between HTLC CLTV timeout and the current block height, for the terminal case (C).

Note that a node is at risk if it accepts an HTLC in one channel and offers an HTLC in another channel with too small of a difference between the CLTV timeouts. For this reason, the cltv_expiry_delta for the outgoing channel is used as the delta across a node.

The worst-case number of blocks between outgoing and incoming HTLC resolution can be derived, given a few assumptions:

  • a worst-case reorganization depth R blocks
  • a grace-period G blocks after HTLC timeout before giving up on an unresponsive peer and dropping to chain
  • a number of blocks S between transaction broadcast and the transaction being included in a block

The worst case is for a forwarding node (B) that takes the longest possible time to spot the outgoing HTLC fulfillment and also takes the longest possible time to redeem it on-chain:

  1. The B->C HTLC times out at block N, and B waits G blocks until it gives up waiting for C. B or C commits to the blockchain, and B spends HTLC, which takes S blocks to be included.
  2. Bad case: C wins the race (just) and fulfills the HTLC, B only sees that transaction when it sees block N+G+S+1.
  3. Worst case: There's reorganization R deep in which C wins and fulfills. B only sees transaction at N+G+S+R.
  4. B now needs to fulfill the incoming A->B HTLC, but A is unresponsive: B waits G more blocks before giving up waiting for A. A or B commits to the blockchain.
  5. Bad case: B sees A's commitment transaction in block N+G+S+R+G+1 and has to spend the HTLC output, which takes S blocks to be mined.
  6. Worst case: there's another reorganization R deep which A uses to spend the commitment transaction, so B sees A's commitment transaction in block N+G+S+R+G+R and has to spend the HTLC output, which takes S blocks to be mined.
  7. B's HTLC spend needs to be at least R deep before it times out, otherwise another reorganization could allow A to timeout the transaction.

Thus, the worst case is 3R+2G+2S, assuming R is at least 1. Note that the chances of three reorganizations in which the other node wins all of them is low for R of 2 or more. Since high fees are used (and HTLC spends can use almost arbitrary fees), S should be small during normal operation; although, given that block times are irregular, empty blocks still occur, fees may vary greatly, and the fees cannot be bumped on HTLC transactions, S=12 should be considered a minimum. S is also the parameter that may vary the most under attack, so a higher value may be desirable when non-negligible amounts are at risk. The grace period G can be low (1 or 2), as nodes are required to timeout or fulfill as soon as possible; but if G is too low it increases the risk of unnecessary channel closure due to networking delays.

There are four values that need be derived:

  1. the cltv_expiry_delta for channels, 3R+2G+2S: if in doubt, a cltv_expiry_delta of at least 34 is reasonable (R=2, G=2, S=12).

  2. the deadline for offered HTLCs: the deadline after which the channel has to be failed and timed out on-chain. This is G blocks after the HTLC's cltv_expiry: 1 or 2 blocks is reasonable.

  3. the deadline for received HTLCs this node has fulfilled: the deadline after which the channel has to be failed and the HTLC fulfilled on-chain before its cltv_expiry. See steps 4-7 above, which imply a deadline of 2R+G+S blocks before cltv_expiry: 18 blocks is reasonable.

  4. the minimum cltv_expiry accepted for terminal payments: the worst case for the terminal node C is 2R+G+S blocks (as, again, steps 1-3 above don't apply). The default in BOLT #11 is 18, which matches this calculation.

Requirements

An offering node:

  • MUST estimate a timeout deadline for each HTLC it offers.
  • MUST NOT offer an HTLC with a timeout deadline before its cltv_expiry.
  • if an HTLC which it offered is in either node's current commitment transaction, AND is past this timeout deadline:
    • SHOULD send an error to the receiving peer (if connected).
    • MUST fail the channel.

A fulfilling node:

  • for each HTLC it is attempting to fulfill:
    • MUST estimate a fulfillment deadline.
  • MUST fail (and not forward) an HTLC whose fulfillment deadline is already past.
  • if an HTLC it has fulfilled is in either node's current commitment transaction, AND is past this fulfillment deadline:
    • SHOULD send an error to the offering peer (if connected).
    • MUST fail the channel.

Adding an HTLC: update_add_htlc

Either node can send update_add_htlc to offer an HTLC to the other, which is redeemable in return for a payment preimage. Amounts are in millisatoshi, though on-chain enforcement is only possible for whole satoshi amounts greater than the dust limit (in commitment transactions these are rounded down as specified in BOLT #3).

The format of the onion_routing_packet portion, which indicates where the payment is destined, is described in BOLT #4.

  1. type: 128 (update_add_htlc)
  2. data:
    • [channel_id:channel_id]
    • [u64:id]
    • [u64:amount_msat]
    • [sha256:payment_hash]
    • [u32:cltv_expiry]
    • [1366*byte:onion_routing_packet]

Requirements

A sending node:

  • if it is responsible for paying the Bitcoin fee:
    • MUST NOT offer amount_msat if, after adding that HTLC to its commitment transaction, it cannot pay the fee for either the local or remote commitment transaction at the current feerate_per_kw while maintaining its channel reserve (see Updating Fees).
    • if option_anchors applies to this commitment transaction and the sending node is the funder:
      • MUST be able to additionally pay for to_local_anchor and to_remote_anchor above its reserve.
    • SHOULD NOT offer amount_msat if, after adding that HTLC to its commitment transaction, its remaining balance doesn't allow it to pay the commitment transaction fee when receiving or sending a future additional non-dust HTLC while maintaining its channel reserve. It is recommended that this "fee spike buffer" can handle twice the current feerate_per_kw to ensure predictability between implementations.
  • if it is not responsible for paying the Bitcoin fee:
    • SHOULD NOT offer amount_msat if, once the remote node adds that HTLC to its commitment transaction, it cannot pay the fee for the updated local or remote transaction at the current feerate_per_kw while maintaining its channel reserve.
  • MUST offer amount_msat greater than 0.
  • MUST NOT offer amount_msat below the receiving node's htlc_minimum_msat
  • MUST set cltv_expiry less than 500000000.
  • if result would be offering more than the remote's max_accepted_htlcs HTLCs, in the remote commitment transaction:
    • MUST NOT add an HTLC.
  • if the sum of total offered HTLCs would exceed the remote's max_htlc_value_in_flight_msat:
    • MUST NOT add an HTLC.
  • for the first HTLC it offers:
    • MUST set id to 0.
  • MUST increase the value of id by 1 for each successive offer.

id MUST NOT be reset to 0 after the update is complete (i.e. after revoke_and_ack has been received). It MUST continue incrementing instead.

A receiving node:

  • receiving an amount_msat equal to 0, OR less than its own htlc_minimum_msat:
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
  • receiving an amount_msat that the sending node cannot afford at the current feerate_per_kw (while maintaining its channel reserve and any to_local_anchor and to_remote_anchor costs):
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
  • if a sending node adds more than receiver max_accepted_htlcs HTLCs to its local commitment transaction, OR adds more than receiver max_htlc_value_in_flight_msat worth of offered HTLCs to its local commitment transaction:
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
  • if sending node sets cltv_expiry to greater or equal to 500000000:
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
  • MUST allow multiple HTLCs with the same payment_hash.
  • if the sender did not previously acknowledge the commitment of that HTLC:
    • MUST ignore a repeated id value after a reconnection.
  • if other id violations occur:
    • MAY send a warning and close the connection, or send an error and fail the channel.

The onion_routing_packet contains an obfuscated list of hops and instructions for each hop along the path. It commits to the HTLC by setting the payment_hash as associated data, i.e. includes the payment_hash in the computation of HMACs. This prevents replay attacks that would reuse a previous onion_routing_packet with a different payment_hash.

Rationale

Invalid amounts are a clear protocol violation and indicate a breakdown.

If a node did not accept multiple HTLCs with the same payment hash, an attacker could probe to see if a node had an existing HTLC. This requirement, to deal with duplicates, leads to the use of a separate identifier; it's assumed a 64-bit counter never wraps.

Retransmissions of unacknowledged updates are explicitly allowed for reconnection purposes; allowing them at other times simplifies the recipient code (though strict checking may help debugging).

max_accepted_htlcs is limited to 483 to ensure that, even if both sides send the maximum number of HTLCs, the commitment_signed message will still be under the maximum message size. It also ensures that a single penalty transaction can spend the entire commitment transaction, as calculated in BOLT #5.

cltv_expiry values equal to or greater than 500000000 would indicate a time in seconds, and the protocol only supports an expiry in blocks.

The node responsible for paying the Bitcoin fee should maintain a "fee spike buffer" on top of its reserve to accommodate a future fee increase. Without this buffer, the node responsible for paying the Bitcoin fee may reach a state where it is unable to send or receive any non-dust HTLC while maintaining its channel reserve (because of the increased weight of the commitment transaction), resulting in a degraded channel. See #728 for more details.

Removing an HTLC: update_fulfill_htlc, update_fail_htlc, and update_fail_malformed_htlc

For simplicity, a node can only remove HTLCs added by the other node. There are four reasons for removing an HTLC: the payment preimage is supplied, it has timed out, it has failed to route, or it is malformed.

To supply the preimage:

  1. type: 130 (update_fulfill_htlc)
  2. data:
    • [channel_id:channel_id]
    • [u64:id]
    • [32*byte:payment_preimage]

For a timed out or route-failed HTLC:

  1. type: 131 (update_fail_htlc)
  2. data:
    • [channel_id:channel_id]
    • [u64:id]
    • [u16:len]
    • [len*byte:reason]

The reason field is an opaque encrypted blob for the benefit of the original HTLC initiator, as defined in BOLT #4; however, there's a special malformed failure variant for the case where the peer couldn't parse it: in this case the current node instead takes action, encrypting it into a update_fail_htlc for relaying.

For an unparsable HTLC:

  1. type: 135 (update_fail_malformed_htlc)
  2. data:
    • [channel_id:channel_id]
    • [u64:id]
    • [sha256:sha256_of_onion]
    • [u16:failure_code]

Requirements

A node:

  • SHOULD remove an HTLC as soon as it can.
  • SHOULD fail an HTLC which has timed out.
  • until the corresponding HTLC is irrevocably committed in both sides' commitment transactions:
    • MUST NOT send an update_fulfill_htlc, update_fail_htlc, or update_fail_malformed_htlc.

A receiving node:

  • if the id does not correspond to an HTLC in its current commitment transaction:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if the payment_preimage value in update_fulfill_htlc doesn't SHA256 hash to the corresponding HTLC payment_hash:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if the BADONION bit in failure_code is not set for update_fail_malformed_htlc:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if the sha256_of_onion in update_fail_malformed_htlc doesn't match the onion it sent:
    • MAY retry or choose an alternate error response.
  • otherwise, a receiving node which has an outgoing HTLC canceled by update_fail_malformed_htlc:
    • MUST return an error in the update_fail_htlc sent to the link which originally sent the HTLC, using the failure_code given and setting the data to sha256_of_onion.

Rationale

A node that doesn't time out HTLCs risks channel failure (see cltv_expiry_delta Selection).

A node that sends update_fulfill_htlc, before the sender, is also committed to the HTLC and risks losing funds.

If the onion is malformed, the upstream node won't be able to extract the shared key to generate a response — hence the special failure message, which makes this node do it.

The node can check that the SHA256 that the upstream is complaining about does match the onion it sent, which may allow it to detect random bit errors. However, without re-checking the actual encrypted packet sent, it won't know whether the error was its own or the remote's; so such detection is left as an option.

Committing Updates So Far: commitment_signed

When a node has changes for the remote commitment, it can apply them, sign the resulting transaction (as defined in BOLT #3), and send a commitment_signed message.

  1. type: 132 (commitment_signed)
  2. data:
    • [channel_id:channel_id]
    • [signature:signature]
    • [u16:num_htlcs]
    • [num_htlcs*signature:htlc_signature]

Requirements

A sending node:

  • MUST NOT send a commitment_signed message that does not include any updates.
  • MAY send a commitment_signed message that only alters the fee.
  • MAY send a commitment_signed message that doesn't change the commitment transaction aside from the new revocation number (due to dust, identical HTLC replacement, or insignificant or multiple fee changes).
  • MUST include one htlc_signature for every HTLC transaction corresponding to the ordering of the commitment transaction (see BOLT #3).
  • if it has not recently received a message from the remote node:
    • SHOULD use ping and await the reply pong before sending commitment_signed.

A receiving node:

  • once all pending updates are applied:
    • if signature is not valid for its local commitment transaction OR non-compliant with LOW-S-standard rule LOWS:
      • MUST send a warning and close the connection, or send an error and fail the channel.
    • if num_htlcs is not equal to the number of HTLC outputs in the local commitment transaction:
      • MUST send a warning and close the connection, or send an error and fail the channel.
  • if any htlc_signature is not valid for the corresponding HTLC transaction OR non-compliant with LOW-S-standard rule LOWS:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • MUST respond with a revoke_and_ack message.

Rationale

There's little point offering spam updates: it implies a bug.

The num_htlcs field is redundant, but makes the packet length check fully self-contained.

The recommendation to require recent messages recognizes the reality that networks are unreliable: nodes might not realize their peers are offline until after sending commitment_signed. Once commitment_signed is sent, the sender considers itself bound to those HTLCs, and cannot fail the related incoming HTLCs until the output HTLCs are fully resolved.

Note that the htlc_signature implicitly enforces the time-lock mechanism in the case of offered HTLCs being timed out or received HTLCs being spent. This is done to reduce fees by creating smaller scripts compared to explicitly stating time-locks on HTLC outputs.

The option_anchors allows HTLC transactions to "bring their own fees" by attaching other inputs and outputs, hence the modified signature flags.

Completing the Transition to the Updated State: revoke_and_ack

Once the recipient of commitment_signed checks the signature and knows it has a valid new commitment transaction, it replies with the commitment preimage for the previous commitment transaction in a revoke_and_ack message.

This message also implicitly serves as an acknowledgment of receipt of the commitment_signed, so this is a logical time for the commitment_signed sender to apply (to its own commitment) any pending updates it sent before that commitment_signed.

The description of key derivation is in BOLT #3.

  1. type: 133 (revoke_and_ack)
  2. data:
    • [channel_id:channel_id]
    • [32*byte:per_commitment_secret]
    • [point:next_per_commitment_point]

Requirements

A sending node:

  • MUST set per_commitment_secret to the secret used to generate keys for the previous commitment transaction.
  • MUST set next_per_commitment_point to the values for its next commitment transaction.

A receiving node:

  • if per_commitment_secret is not a valid secret key or does not generate the previous per_commitment_point:
    • MUST send an error and fail the channel.
  • if the per_commitment_secret was not generated by the protocol in BOLT #3:
    • MAY send a warning and close the connection, or send an error and fail the channel.

A node:

  • MUST NOT broadcast old (revoked) commitment transactions,
    • Note: doing so will allow the other node to seize all channel funds.
  • SHOULD NOT sign commitment transactions, unless it's about to broadcast them (due to a failed connection),
    • Note: this is to reduce the above risk.

Updating Fees: update_fee

An update_fee message is sent by the node which is paying the Bitcoin fee. Like any update, it's first committed to the receiver's commitment transaction and then (once acknowledged) committed to the sender's. Unlike an HTLC, update_fee is never closed but simply replaced.

There is a possibility of a race, as the recipient can add new HTLCs before it receives the update_fee. Under this circumstance, the sender may not be able to afford the fee on its own commitment transaction, once the update_fee is finally acknowledged by the recipient. In this case, the fee will be less than the fee rate, as described in BOLT #3.

The exact calculation used for deriving the fee from the fee rate is given in BOLT #3.

  1. type: 134 (update_fee)
  2. data:
    • [channel_id:channel_id]
    • [u32:feerate_per_kw]

Requirements

The node responsible for paying the Bitcoin fee:

  • SHOULD send update_fee to ensure the current fee rate is sufficient (by a significant margin) for timely processing of the commitment transaction.

The node not responsible for paying the Bitcoin fee:

  • MUST NOT send update_fee.

A receiving node:

  • if the update_fee is too low for timely processing, OR is unreasonably large:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if the sender is not responsible for paying the Bitcoin fee:
    • MUST send a warning and close the connection, or send an error and fail the channel.
  • if the sender cannot afford the new fee rate on the receiving node's current commitment transaction:
    • SHOULD send a warning and close the connection, or send an error and fail the channel.
      • but MAY delay this check until the update_fee is committed.

Rationale

Bitcoin fees are required for unilateral closes to be effective. With option_anchors, feerate_per_kw is not as critical anymore to guarantee confirmation as it was in the legacy commitment format, but it still needs to be enough to be able to enter the mempool (satisfy min relay fee and mempool min fee).

For the legacy commitment format, there is no general method for the broadcasting node to use child-pays-for-parent to increase its effective fee.

Given the variance in fees, and the fact that the transaction may be spent in the future, it's a good idea for the fee payer to keep a good margin (say 5x the expected fee requirement) for legacy commitment txes; but, due to differing methods of fee estimation, an exact value is not specified.

Since the fees are currently one-sided (the party which requested the channel creation always pays the fees for the commitment transaction), it's simplest to only allow it to set fee levels; however, as the same fee rate applies to HTLC transactions, the receiving node must also care about the reasonableness of the fee.

Message Retransmission

Because communication transports are unreliable, and may need to be re-established from time to time, the design of the transport has been explicitly separated from the protocol.

Nonetheless, it's assumed our transport is ordered and reliable. Reconnection introduces doubt as to what has been received, so there are explicit acknowledgments at that point.

This is fairly straightforward in the case of channel establishment and close, where messages have an explicit order, but during normal operation, acknowledgments of updates are delayed until the commitment_signed / revoke_and_ack exchange; so it cannot be assumed that the updates have been received. This also means that the receiving node only needs to store updates upon receipt of commitment_signed.

Note that messages described in BOLT #7 are independent of particular channels; their transmission requirements are covered there, and besides being transmitted after init (as all messages are), they are independent of requirements here.

  1. type: 136 (channel_reestablish)
  2. data:
    • [channel_id:channel_id]
    • [u64:next_commitment_number]
    • [u64:next_revocation_number]
    • [32*byte:your_last_per_commitment_secret]
    • [point:my_current_per_commitment_point]

next_commitment_number: A commitment number is a 48-bit incrementing counter for each commitment transaction; counters are independent for each peer in the channel and start at 0. They're only explicitly relayed to the other node in the case of re-establishment, otherwise they are implicit.

Requirements

A funding node:

  • upon disconnection:
    • if it has broadcast the funding transaction:
      • MUST remember the channel for reconnection.
    • otherwise:
      • SHOULD NOT remember the channel for reconnection.

A non-funding node:

  • upon disconnection:
    • if it has sent the funding_signed message:
      • MUST remember the channel for reconnection.
    • otherwise:
      • SHOULD NOT remember the channel for reconnection.

A node:

  • MUST handle continuation of a previous channel on a new encrypted transport.
  • upon disconnection:
    • MUST reverse any uncommitted updates sent by the other side (i.e. all messages beginning with update_ for which no commitment_signed has been received).
      • Note: a node MAY have already used the payment_preimage value from the update_fulfill_htlc, so the effects of update_fulfill_htlc are not completely reversed.
  • upon reconnection:
    • if a channel is in an error state:
      • SHOULD retransmit the error packet and ignore any other packets for that channel.
    • otherwise:
      • MUST transmit channel_reestablish for each channel.
      • MUST wait to receive the other node's channel_reestablish message before sending any other messages for that channel.

The sending node:

  • MUST set next_commitment_number to the commitment number of the next commitment_signed it expects to receive.
  • MUST set next_revocation_number to the commitment number of the next revoke_and_ack message it expects to receive.
  • if option_static_remotekey applies to the commitment transaction:
    • MUST set my_current_per_commitment_point to a valid point.
  • otherwise:
    • MUST set my_current_per_commitment_point to its commitment point for the last signed commitment it received from its channel peer (i.e. the commitment_point corresponding to the commitment transaction the sender would use to unilaterally close).
  • if next_revocation_number equals 0:
    • MUST set your_last_per_commitment_secret to all zeroes
  • otherwise:
    • MUST set your_last_per_commitment_secret to the last per_commitment_secret it received

A node:

  • if next_commitment_number is 1 in both the channel_reestablish it sent and received:
    • MUST retransmit funding_locked.
  • otherwise:
    • MUST NOT retransmit funding_locked.
  • upon reconnection:
    • MUST ignore any redundant funding_locked it receives.
  • if next_commitment_number is equal to the commitment number of the last commitment_signed message the receiving node has sent:
    • MUST reuse the same commitment number for its next commitment_signed.
  • otherwise:
    • if next_commitment_number is not 1 greater than the commitment number of the last commitment_signed message the receiving node has sent:
      • SHOULD send an error and fail the channel.
    • if it has not sent commitment_signed, AND next_commitment_number is not equal to 1:
      • SHOULD send an error and fail the channel.
  • if next_revocation_number is equal to the commitment number of the last revoke_and_ack the receiving node sent, AND the receiving node hasn't already received a closing_signed:
    • MUST re-send the revoke_and_ack.
    • if it has previously sent a commitment_signed that needs to be retransmitted:
      • MUST retransmit revoke_and_ack and commitment_signed in the same relative order as initially transmitted.
  • otherwise:
    • if next_revocation_number is not equal to 1 greater than the commitment number of the last revoke_and_ack the receiving node has sent:
      • SHOULD send an error and fail the channel.
    • if it has not sent revoke_and_ack, AND next_revocation_number is not equal to 0:
      • SHOULD send an error and fail the channel.

A receiving node:

  • if option_static_remotekey applies to the commitment transaction:
    • if next_revocation_number is greater than expected above, AND your_last_per_commitment_secret is correct for that next_revocation_number minus 1:
      • MUST NOT broadcast its commitment transaction.
      • SHOULD send an error to request the peer to fail the channel.
    • otherwise:
      • if your_last_per_commitment_secret does not match the expected values:
        • SHOULD send an error and fail the channel.
  • otherwise, if it supports option_data_loss_protect:
    • if next_revocation_number is greater than expected above, AND your_last_per_commitment_secret is correct for that next_revocation_number minus 1:
      • MUST NOT broadcast its commitment transaction.
      • SHOULD send an error to request the peer to fail the channel.
      • SHOULD store my_current_per_commitment_point to retrieve funds should the sending node broadcast its commitment transaction on-chain.
    • otherwise (your_last_per_commitment_secret or my_current_per_commitment_point do not match the expected values):
      • SHOULD send an error and fail the channel.

A node:

  • MUST NOT assume that previously-transmitted messages were lost,
    • if it has sent a previous commitment_signed message:
      • MUST handle the case where the corresponding commitment transaction is broadcast at any time by the other side,
        • Note: this is particularly important if the node does not simply retransmit the exact update_ messages as previously sent.
  • upon reconnection:
    • if it has sent a previous shutdown:
      • MUST retransmit shutdown.

Rationale

The requirements above ensure that the opening phase is nearly atomic: if it doesn't complete, it starts again. The only exception is if the funding_signed message is sent but not received. In this case, the funder will forget the channel, and presumably open a new one upon reconnection; meanwhile, the other node will eventually forget the original channel, due to never receiving funding_locked or seeing the funding transaction on-chain.

There's no acknowledgment for error, so if a reconnect occurs it's polite to retransmit before disconnecting again; however, it's not a MUST, because there are also occasions where a node can simply forget the channel altogether.

closing_signed also has no acknowledgment so must be retransmitted upon reconnection (though negotiation restarts on reconnection, so it needs not be an exact retransmission). The only acknowledgment for shutdown is closing_signed, so one or the other needs to be retransmitted.

The handling of updates is similarly atomic: if the commit is not acknowledged (or wasn't sent) the updates are re-sent. However, it's not insisted they be identical: they could be in a different order, involve different fees, or even be missing HTLCs which are now too old to be added. Requiring they be identical would effectively mean a write to disk by the sender upon each transmission, whereas the scheme here encourages a single persistent write to disk for each commitment_signed sent or received. But if you need to retransmit both a commitment_signed and a revoke_and_ack, the relative order of these two must be preserved, otherwise it will lead to a channel closure.

A re-transmittal of revoke_and_ack should never be asked for after a closing_signed has been received, since that would imply a shutdown has been completed — which can only occur after the revoke_and_ack has been received by the remote node.

Note that the next_commitment_number starts at 1, since commitment number 0 is created during opening. next_revocation_number will be 0 until the commitment_signed for commitment number 1 is send and then the revocation for commitment number 0 is received.

funding_locked is implicitly acknowledged by the start of normal operation, which is known to have begun after a commitment_signed has been received — hence, the test for a next_commitment_number greater than 1.

A previous draft insisted that the funder "MUST remember ...if it has broadcast the funding transaction, otherwise it MUST NOT": this was in fact an impossible requirement. A node must either firstly commit to disk and secondly broadcast the transaction or vice versa. The new language reflects this reality: it's surely better to remember a channel which hasn't been broadcast than to forget one which has! Similarly, for the fundee's funding_signed message: it's better to remember a channel that never opens (and times out) than to let the funder open it while the fundee has forgotten it.

option_data_loss_protect was added to allow a node, which has somehow fallen behind (e.g. has been restored from old backup), to detect that it has fallen behind. A fallen-behind node must know it cannot broadcast its current commitment transaction — which would lead to total loss of funds — as the remote node can prove it knows the revocation preimage. The error returned by the fallen-behind node should make the other node drop its current commitment transaction to the chain. The other node should wait for that error to give the fallen-behind node an opportunity to fix its state first (e.g by restarting with a different backup). If the fallen-behind node doesn't have the latest backup, this will, at least, allow it to recover non-HTLC funds, if the my_current_per_commitment_point is valid. However, this also means the fallen-behind node has revealed this fact (though not provably: it could be lying), and the other node could use this to broadcast a previous state.

option_static_remotekey removes the changing to_remote key, so the my_current_per_commitment_point is unnecessary and thus ignored (for parsing simplicity, it remains and must be a valid point, however), but the disclosure of previous secret still allows fall-behind detection. An implementation can offer both, however, and fall back to the option_data_loss_protect behavior if option_static_remotekey is not negotiated.

Authors

[ FIXME: Insert Author List ]

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.