MIME-Version: 1.0
References: <CAFXO6=+cHyQKM2n9yn4EhwLZO+AUB0ZD81qWPxmpN27rjUoU3w@mail.gmail.com>
 <CALZpt+HpvmEHUEOgye34T6pVQ+wnKKn-_8cTJTQXYQb9t1jOTA@mail.gmail.com>
 <CAFXO6=JzsYgiXJE2geSKMpfgPo+GGNX_+Pw0JQx1QQxAfhCdBQ@mail.gmail.com>
 <CALZpt+HQpdrebhWGXv_yLqiSCB5Ur71L1K13bd7w5TZb9DwJEQ@mail.gmail.com>
 <CAFXO6=Lvcr7Pwn_ZD1CZohYUFKY-cC5sGRxdTOiP2MgnTvFnYA@mail.gmail.com>
 <CALZpt+FFSk-+BBxu7SSdjw580UCFfkdo1DTa1Yj9K81M4E1vPg@mail.gmail.com>
In-Reply-To: <CALZpt+FFSk-+BBxu7SSdjw580UCFfkdo1DTa1Yj9K81M4E1vPg@mail.gmail.com>
From: Bastien TEINTURIER <bastien@acinq.fr>
Date: Mon, 27 Sep 2021 09:15:18 +0200
Message-ID: <CACdvm3Pgia5mn60HmZmi4t0U5coc13WBuDkh+QgQBAHZz+Ufow@mail.gmail.com>
To: Antoine Riard <antoine.riard@gmail.com>, 
 Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
Content-Type: multipart/alternative; boundary="000000000000a309eb05ccf4dc02"
Subject: Re: [bitcoin-dev] Proposal: Package Mempool Accept and Package RBF
Precedence: list

--000000000000a309eb05ccf4dc02
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

>
> I think we could restrain package acceptance to only confirmed inputs for
> now and revisit later this point ? For LN-anchor, you can assume that the
> fee-bumping UTXO feeding the CPFP is already
> confirmed. Or are there currently-deployed use-cases which would benefit
> from your proposed Rule #2 ?
>

I think constraining package acceptance to only confirmed inputs
is very limiting and quite dangerous for L2 protocols.

In the case of LN, an attacker can game this and heavily restrict
your RBF attempts if you're only allowed to use confirmed inputs
and have many channels (and a limited number of confirmed inputs).
Otherwise you'll need node operators to pre-emptively split their
utxos into many small utxos just for fee bumping, which is inefficient...

Bastien

Le lun. 27 sept. 2021 =C3=A0 00:27, Antoine Riard via bitcoin-dev <
bitcoin-dev@lists.linuxfoundation.org> a =C3=A9crit :

> Hi Gloria,
>
> Thanks for your answers,
>
> > In summary, it seems that the decisions that might still need
> > attention/input from devs on this mailing list are:
> > 1. Whether we should start with multiple-parent-1-child or
> 1-parent-1-child.
> > 2. Whether it's ok to require that the child not have conflicts with
> > mempool transactions.
>
> Yes 1) it would be good to have inputs of more potential users of package
> acceptance . And 2) I think it's more a matter of clearer wording of the
> proposal.
>
> However, see my final point on the relaxation around "unconfirmed inputs"
> which might in fact alter our current block construction strategy.
>
> > Right, the fact that we essentially always choose the first-seen witnes=
s
> is
> > an unfortunate limitation that exists already. Adding package mempool
> > accept doesn't worsen this, but the procedure in the future is to repla=
ce
> > the witness when it makes sense economically. We can also add logic to
> > allow package feerate to pay for witness replacements as well. This is
> > pretty far into the future, though.
>
> Yes I agree package mempool doesn't worsen this. And it's not an issue fo=
r
> current LN as you can't significantly inflate a spending witness for the
> 2-of-2 funding output.
> However, it might be an issue for multi-party protocol where the spending
> script has alternative branches with asymmetric valid witness weights.
> Taproot should ease that kind of script so hopefully we would deploy
> wtxid-replacement not too far in the future.
>
> > I could be misunderstanding, but an attacker wouldn't be able to
> > batch-attack like this. Alice's package only conflicts with A' + D', no=
t
> A'
> > + B' + C' + D'. She only needs to pay for evicting 2 transactions.
>
> Yeah I can be clearer, I think you have 2 pinning attacks scenarios to
> consider.
>
> In LN, if you're trying to confirm a commitment transaction to time-out o=
r
> claim on-chain a HTLC and the timelock is near-expiration, you should be
> ready to pay in commitment+2nd-stage HTLC transaction fees as much as the
> value offered by the HTLC.
>
> Following this security assumption, an attacker can exploit it by
> targeting together commitment transactions from different channels by
> blocking them under a high-fee child, of which the fee value
> is equal to the top-value HTLC + 1. Victims's fee-bumping logics won't
> overbid as it's not worthy to offer fees beyond their competed HTLCs. Apa=
rt
> from observing mempools state, victims can't learn they're targeted by th=
e
> same attacker.
>
> To draw from the aforementioned topology, Mallory broadcasts A' + B' + C'
> + D', where A' conflicts with Alice's P1, B' conflicts with Bob's P2, C'
> conflicts with Caroll's P3. Let's assume P1 is confirming the top-value
> HTLC of the set. If D' fees is higher than P1 + 1, it won't be rational f=
or
> Alice or Bob or Caroll to keep offering competing feerates. Mallory will =
be
> at loss on stealing P1, as she has paid more in fees but will realize a
> gain on P2+P3.
>
> In this model, Alice is allowed to evict those 2 transactions (A' + D')
> but as she is economically-bounded she won't succeed.
>
> Mallory is maliciously exploiting RBF rule 3 on absolute fee. I think thi=
s
> 1st pinning scenario is correct and "lucractive" when you sum the global
> gain/loss.
>
> There is a 2nd attack scenario where A + B + C + D, where D is the child
> of A,B,C. All those transactions are honestly issued by Alice. Once A + B=
 +
> C + D are propagated in network mempools, Mallory is able to replace A + =
D
> with  A' + D' where D' is paying a higher fee. This package A' + D' will
> confirm soon if D feerate was compelling but Mallory succeeds in delaying
> the confirmation
> of B + C for one or more blocks. As B + C are pre-signed commitments with
> a low-fee rate they won't confirm without Alice issuing a new child E.
> Mallory can repeat the same trick by broadcasting
> B' + E' and delay again the confirmation of C.
>
> If the remaining package pending HTLC has a higher-value than all the
> malicious fees over-bid, Mallory should realize a gain. With this 2nd
> pinning attack, the malicious entity buys confirmation delay of your
> packaged-together commitments.
>
> Assuming those attacks are correct, I'm leaning towards being conservativ=
e
> with the LDK broadcast backend. Though once again, other L2 devs have
> likely other use-cases and opinions :)
>
> >  B' only needs to pay for itself in this case.
>
> Yes I think it's a nice discount when UTXO is single-owned. In the contex=
t
> of shared-owned UTXO (e.g LN), you might not if there is an in-mempool
> package already spending the UTXO and have to assume the worst-case
> scenario. I.e have B' committing enough fee to pay for A' replacement
> bandwidth. I think we can't do that much for this case...
>
> > If a package meets feerate requirements as a
> package, the parents in the transaction are allowed to replace-by-fee
> mempool transactions. The child cannot replace mempool transactions."
>
> I agree with the Mallory-vs-Alice case. Though if Alice broadcasts A+B' t=
o
> replace A+B because the first broadcast isn't satisfying anymore due to
> mempool spikes ? Assuming B' fees is enough, I think that case as child B=
'
> replacing in-mempool transaction B. Which I understand going against  "Th=
e
> child cannot replace mempool transactions".
>
> Maybe wording could be a bit clearer ?
>
> > While it would be nice to have full RBF, malleability of the child won'=
t
> > block RBF here. If we're trying to replace A', we only require that A'
> > signals replaceability, and don't mind if its child doesn't.
>
> Yes, it sounds good.
>
> > Yes, A+C+D pays 2500sat more in fees, but it is also 1000vB larger. A
> miner
> > should prefer to utilize their block space more effectively.
>
> If your mempool is empty and only composed of A+C+D or A+B, I think takin=
g
> A+C+D is the most efficient block construction you can come up with as a
> miner ?
>
> > No, because we don't use that model.
>
> Can you describe what miner model we are using ? Like the block
> construction strategy implemented by `addPackagesTxs` or also encompassin=
g
> our current mempool acceptance policy, which I think rely on absolute fee
> over ancestor score in case of replacement ?
>
> I think this point is worthy to discuss as otherwise we might downgrade
> the efficiency of our current block construction strategy in periods of
> near-empty mempools. A knowledge which could be discreetly leveraged by a
> miner to gain an advantage on the rest of the mining ecosystem.
>
> Note, I think we *might* have to go in this direction if we want to
> replace replace-by-fee by replace-by-feerate or replace-by-ancestor and
> solve in-depth pinning attacks. Though if we do so,
> IMO we would need more thoughts.
>
> I think we could restrain package acceptance to only confirmed inputs for
> now and revisit later this point ? For LN-anchor, you can assume that the
> fee-bumping UTXO feeding the CPFP is already
> confirmed. Or are there currently-deployed use-cases which would benefit
> from your proposed Rule #2 ?
>
> Antoine
>
> Le jeu. 23 sept. 2021 =C3=A0 11:36, Gloria Zhao <gloriajzhao@gmail.com> a
> =C3=A9crit :
>
>> Hi Antoine,
>>
>> Thanks as always for your input. I'm glad we agree on so much!
>>
>> In summary, it seems that the decisions that might still need
>> attention/input from devs on this mailing list are:
>> 1. Whether we should start with multiple-parent-1-child or
>> 1-parent-1-child.
>> 2. Whether it's ok to require that the child not have conflicts with
>> mempool transactions.
>>
>> Responding to your comments...
>>
>> > IIUC, you have package A+B, during the dedup phase early in
>> `AcceptMultipleTransactions` if you observe same-txid-different-wtixd A'
>> and A' is higher feerate than A, you trim A and replace by A' ?
>>
>> > I think this approach is safe, the one who appears unsafe to me is whe=
n
>> A' has a _lower_ feerate, even if A' is already accepted by our mempool =
?
>> In that case iirc that would be a pinning.
>>
>> Right, the fact that we essentially always choose the first-seen witness
>> is an unfortunate limitation that exists already. Adding package mempool
>> accept doesn't worsen this, but the procedure in the future is to replac=
e
>> the witness when it makes sense economically. We can also add logic to
>> allow package feerate to pay for witness replacements as well. This is
>> pretty far into the future, though.
>>
>> > It sounds uneconomical for an attacker but I think it's not when you
>> consider than you can "batch" attack against multiple honest
>> counterparties. E.g, Mallory broadcast A' + B' + C' + D' where A' confli=
cts
>> with Alice's honest package P1, B' conflicts with Bob's honest package P=
2,
>> C' conflicts with Caroll's honest package P3. And D' is a high-fee child=
 of
>> A' + B' + C'.
>>
>> > If D' is higher-fee than P1 or P2 or P3 but inferior to the sum of
>> HTLCs confirmed by P1+P2+P3, I think it's lucrative for the attacker ?
>>
>> I could be misunderstanding, but an attacker wouldn't be able to
>> batch-attack like this. Alice's package only conflicts with A' + D', not=
 A'
>> + B' + C' + D'. She only needs to pay for evicting 2 transactions.
>>
>> > Do we assume that broadcasted packages are "honest" by default and tha=
t
>> the parent(s) always need the child to pass the fee checks, that way sav=
ing
>> the processing of individual transactions which are expected to fail in =
99%
>> of cases or more ad hoc composition of packages at relay ?
>> > I think this point is quite dependent on the p2p packages format/logic
>> we'll end up on and that we should feel free to revisit it later ?
>>
>> I think it's the opposite; there's no way for us to assume that p2p
>> packages will be "honest." I'd like to have two things before we expose =
on
>> P2P: (1) ensure that the amount of resources potentially allocated for
>> package validation isn't disproportionately higher than that of single
>> transaction validation and (2) only use package validation when we're
>> unsatisifed with the single validation result, e.g. we might get better
>> fees.
>> Yes, let's revisit this later :)
>>
>>  > Yes, if you receive A+B, and A is already in-mempoo, I agree you can
>> discard its feerate as B should pay for all fees checked on its own. Whe=
re
>> I'm unclear is when you have in-mempool A+B and receive A+B'. Should B'
>> have a fee high enough to cover the bandwidth penalty replacement
>> (`PaysForRBF`, 2nd check) of both A+B' or only B' ?
>>
>>  B' only needs to pay for itself in this case.
>>
>> > > Do we want the child to be able to replace mempool transactions as
>> well?
>>
>> > If we mean when you have replaceable A+B then A'+B' try to replace wit=
h
>> a higher-feerate ? I think that's exactly the case we need for Lightning=
 as
>> A+B is coming from Alice and A'+B' is coming from Bob :/
>>
>> Let me clarify this because I can see that my wording was ambiguous, and
>> then please let me know if it fits Lightning's needs?
>>
>> In my proposal, I wrote "If a package meets feerate requirements as a
>> package, the parents in the transaction are allowed to replace-by-fee
>> mempool transactions. The child cannot replace mempool transactions." Wh=
at
>> I meant was: the package can replace mempool transactions if any of the
>> parents conflict with mempool transactions. The child cannot not conflic=
t
>> with any mempool transactions.
>> The Lightning use case this attempts to address is: Alice and Mallory ar=
e
>> LN counterparties, and have packages A+B and A'+B', respectively. A and =
A'
>> are their commitment transactions and conflict with each other; they hav=
e
>> shared inputs and different txids.
>> B spends Alice's anchor output from A. B' spends Mallory's anchor output
>> from A'. Thus, B and B' do not conflict with each other.
>> Alice can broadcast her package, A+B, to replace Mallory's package,
>> A'+B', since B doesn't conflict with the mempool.
>>
>> Would this be ok?
>>
>> > The second option, a child of A', In the LN case I think the CPFP is
>> attached on one's anchor output.
>>
>> While it would be nice to have full RBF, malleability of the child won't
>> block RBF here. If we're trying to replace A', we only require that A'
>> signals replaceability, and don't mind if its child doesn't.
>>
>> > > B has an ancestor score of 10sat/vb and D has an
>> > > ancestor score of ~2.9sat/vb. Since D's ancestor score is lower than
>> B's,
>> > > it fails the proposed package RBF Rule #2, so this package would be
>> > > rejected. Does this meet your expectations?
>>
>> > Well what sounds odd to me, in my example, we fail D even if it has a
>> higher-fee than B. Like A+B absolute fees are 2000 sats and A+C+D absolu=
te
>> fees are 4500 sats ?
>>
>> Yes, A+C+D pays 2500sat more in fees, but it is also 1000vB larger. A
>> miner should prefer to utilize their block space more effectively.
>>
>> > Is this compatible with a model where a miner prioritizes absolute fee=
s
>> over ancestor score, in the case that mempools aren't full-enough to
>> fulfill a block ?
>>
>> No, because we don't use that model.
>>
>> Thanks,
>> Gloria
>>
>> On Thu, Sep 23, 2021 at 5:29 AM Antoine Riard <antoine.riard@gmail.com>
>> wrote:
>>
>>> > Correct, if B+C is too low feerate to be accepted, we will reject it.=
 I
>>> > prefer this because it is incentive compatible: A can be mined by
>>> itself,
>>> > so there's no reason to prefer A+B+C instead of A.
>>> > As another way of looking at this, consider the case where we do acce=
pt
>>> > A+B+C and it sits at the "bottom" of our mempool. If our mempool
>>> reaches
>>> > capacity, we evict the lowest descendant feerate transactions, which
>>> are
>>> > B+C in this case. This gives us the same resulting mempool, with A an=
d
>>> not
>>> > B+C.
>>>
>>> I agree here. Doing otherwise, we might evict other transactions mempoo=
l
>>> in `MempoolAccept::Finalize` with a higher-feerate than B+C while those
>>> evicted transactions are the most compelling for block construction.
>>>
>>> I thought at first missing this acceptance requirement would break a
>>> fee-bumping scheme like Parent-Pay-For-Child where a high-fee parent is
>>> attached to a child signed with SIGHASH_ANYONECANPAY but in this case t=
he
>>> child fee is capturing the parent value. I can't think of other fee-bum=
ping
>>> schemes potentially affected. If they do exist I would say they're wron=
g in
>>> their design assumptions.
>>>
>>> > If or when we have witness replacement, the logic is: if the individu=
al
>>> > transaction is enough to replace the mempool one, the replacement wil=
l
>>> > happen during the preceding individual transaction acceptance, and
>>> > deduplication logic will work. Otherwise, we will try to deduplicate =
by
>>> > wtxid, see that we need a package witness replacement, and use the
>>> package
>>> > feerate to evaluate whether this is economically rational.
>>>
>>> IIUC, you have package A+B, during the dedup phase early in
>>> `AcceptMultipleTransactions` if you observe same-txid-different-wtixd A=
'
>>> and A' is higher feerate than A, you trim A and replace by A' ?
>>>
>>> I think this approach is safe, the one who appears unsafe to me is when
>>> A' has a _lower_ feerate, even if A' is already accepted by our mempool=
 ?
>>> In that case iirc that would be a pinning.
>>>
>>> Good to see progress on witness replacement before we see usage of
>>> Taproot tree in the context of multi-party, where a malicious counterpa=
rty
>>> inflates its witness to jam a honest spending.
>>>
>>> (Note, the commit linked currently points nowhere :))
>>>
>>>
>>> > Please note that A may replace A' even if A' has higher fees than A
>>> > individually, because the proposed package RBF utilizes the fees and
>>> size
>>> > of the entire package. This just requires E to pay enough fees,
>>> although
>>> > this can be pretty high if there are also potential B' and C' competi=
ng
>>> > commitment transactions that we don't know about.
>>>
>>> Ah right, if the package acceptance waives `PaysMoreThanConflicts` for
>>> the individual check on A, the honest package should replace the pinnin=
g
>>> attempt. I've not fully parsed the proposed implementation yet.
>>>
>>> Though note, I think it's still unsafe for a Lightning
>>> multi-commitment-broadcast-as-one-package as a malicious A' might have =
an
>>> absolute fee higher than E. It sounds uneconomical for
>>> an attacker but I think it's not when you consider than you can "batch"
>>> attack against multiple honest counterparties. E.g, Mallory broadcast A=
' +
>>> B' + C' + D' where A' conflicts with Alice's honest package P1, B'
>>> conflicts with Bob's honest package P2, C' conflicts with Caroll's hone=
st
>>> package P3. And D' is a high-fee child of A' + B' + C'.
>>>
>>> If D' is higher-fee than P1 or P2 or P3 but inferior to the sum of HTLC=
s
>>> confirmed by P1+P2+P3, I think it's lucrative for the attacker ?
>>>
>>> > So far, my understanding is that multi-parent-1-child is desired for
>>> > batched fee-bumping (
>>> > https://github.com/bitcoin/bitcoin/pull/22674#issuecomment-897951289)
>>> and
>>> > I've also seen your response which I have less context on (
>>> > https://github.com/bitcoin/bitcoin/pull/22674#issuecomment-900352202)=
.
>>> That
>>> > being said, I am happy to create a new proposal for 1 parent + 1 chil=
d
>>> > (which would be slightly simpler) and plan for moving to
>>> > multi-parent-1-child later if that is preferred. I am very interested
>>> in
>>> > hearing feedback on that approach.
>>>
>>> I think batched fee-bumping is okay as long as you don't have
>>> time-sensitive outputs encumbering your commitment transactions. For th=
e
>>> reasons mentioned above, I think that's unsafe.
>>>
>>> What I'm worried about is  L2 developers, potentially not aware about
>>> all the mempool subtleties blurring the difference and always batching
>>> their broadcast by default.
>>>
>>> IMO, a good thing by restraining to 1-parent + 1 child,  we artificiall=
y
>>> constraint L2 design space for now and minimize risks of unsafe usage o=
f
>>> the package API :)
>>>
>>> I think that's a point where it would be relevant to have the opinion o=
f
>>> more L2 devs.
>>>
>>> > I think there is a misunderstanding here - let me describe what I'm
>>> > proposing we'd do in this situation: we'll try individual submission
>>> for A,
>>> > see that it fails due to "insufficient fees." Then, we'll try package
>>> > validation for A+B and use package RBF. If A+B pays enough, it can
>>> still
>>> > replace A'. If A fails for a bad signature, we won't look at B or A+B=
.
>>> Does
>>> > this meet your expectations?
>>>
>>> Yes there was a misunderstanding, I think this approach is correct, it'=
s
>>> more a question of performance. Do we assume that broadcasted packages =
are
>>> "honest" by default and that the parent(s) always need the child to pas=
s
>>> the fee checks, that way saving the processing of individual transactio=
ns
>>> which are expected to fail in 99% of cases or more ad hoc composition o=
f
>>> packages at relay ?
>>>
>>> I think this point is quite dependent on the p2p packages format/logic
>>> we'll end up on and that we should feel free to revisit it later ?
>>>
>>>
>>> > What problem are you trying to solve by the package feerate *after*
>>> dedup
>>> rule ?
>>> > My understanding is that an in-package transaction might be already i=
n
>>> the mempool. Therefore, to compute a correct RBF penalty replacement, t=
he
>>> vsize of this transaction could be discarded lowering the cost of packa=
ge
>>> RBF.
>>>
>>> > I'm proposing that, when a transaction has already been submitted to
>>> > mempool, we would ignore both its fees and vsize when calculating
>>> package
>>> > feerate.
>>>
>>> Yes, if you receive A+B, and A is already in-mempoo, I agree you can
>>> discard its feerate as B should pay for all fees checked on its own. Wh=
ere
>>> I'm unclear is when you have in-mempool A+B and receive A+B'. Should B'
>>> have a fee high enough to cover the bandwidth penalty replacement
>>> (`PaysForRBF`, 2nd check) of both A+B' or only B' ?
>>>
>>> If you have a second-layer like current Lightning, you might have a
>>> counterparty commitment to replace and should always expect to have to =
pay
>>> for parent replacement bandwidth.
>>>
>>> Where a potential discount sounds interesting is when you have an
>>> univoque state on the first-stage of transactions. E.g DLC's funding
>>> transaction which might be CPFP by any participant iirc.
>>>
>>> > Note that, if C' conflicts with C, it also conflicts with D, since D
>>> is a
>>> > descendant of C and would thus need to be evicted along with it.
>>>
>>> Ah once again I think it's a misunderstanding without the code under my
>>> eyes! If we do C' `PreChecks`, solve the conflicts provoked by it, i.e =
mark
>>> for potential eviction D and don't consider it for future conflicts in =
the
>>> rest of the package, I think D' `PreChecks` should be good ?
>>>
>>> > More generally, this example is surprising to me because I didn't thi=
nk
>>> > packages would be used to fee-bump replaceable transactions. Do we
>>> want the
>>> > child to be able to replace mempool transactions as well?
>>>
>>> If we mean when you have replaceable A+B then A'+B' try to replace with
>>> a higher-feerate ? I think that's exactly the case we need for Lightnin=
g as
>>> A+B is coming from Alice and A'+B' is coming from Bob :/
>>>
>>> > I'm not sure what you mean? Let's say we have a package of parent A +
>>> child
>>> > B, where A is supposed to replace a mempool transaction A'. Are you
>>> saying
>>> > that counterparties are able to malleate the package child B, or a
>>> child of
>>> > A'?
>>>
>>> The second option, a child of A', In the LN case I think the CPFP is
>>> attached on one's anchor output.
>>>
>>> I think it's good if we assume the
>>> solve-conflicts-after-parent's`'PreChecks` mentioned above or fixing
>>> inherited signaling or full-rbf ?
>>>
>>> > Sorry, I don't understand what you mean by "preserve the package
>>> > integrity?" Could you elaborate?
>>>
>>> After thinking the relaxation about the "new" unconfirmed input is not
>>> linked to trimming but I would say more to the multi-parent support.
>>>
>>> Let's say you have A+B trying to replace C+D where B is also spending
>>> already in-mempool E. To succeed, you need to waive the no-new-unconfir=
med
>>> input as D isn't spending E.
>>>
>>> So good, I think we agree on the problem description here.
>>>
>>> > I am in agreement with your calculations but unsure if we disagree on
>>> the
>>> > expected outcome. Yes, B has an ancestor score of 10sat/vb and D has =
an
>>> > ancestor score of ~2.9sat/vb. Since D's ancestor score is lower than
>>> B's,
>>> > it fails the proposed package RBF Rule #2, so this package would be
>>> > rejected. Does this meet your expectations?
>>>
>>> Well what sounds odd to me, in my example, we fail D even if it has a
>>> higher-fee than B. Like A+B absolute fees are 2000 sats and A+C+D absol=
ute
>>> fees are 4500 sats ?
>>>
>>> Is this compatible with a model where a miner prioritizes absolute fees
>>> over ancestor score, in the case that mempools aren't full-enough to
>>> fulfill a block ?
>>>
>>> Let me know if I can clarify a point.
>>>
>>> Antoine
>>>
>>> Le lun. 20 sept. 2021 =C3=A0 11:10, Gloria Zhao <gloriajzhao@gmail.com>=
 a
>>> =C3=A9crit :
>>>
>>>>
>>>> Hi Antoine,
>>>>
>>>> First of all, thank you for the thorough review. I appreciate your
>>>> insight on LN requirements.
>>>>
>>>> > IIUC, you have a package A+B+C submitted for acceptance and A is
>>>> already in your mempool. You trim out A from the package and then eval=
uate
>>>> B+C.
>>>>
>>>> > I think this might be an issue if A is the higher-fee element of the
>>>> ABC package. B+C package fees might be under the mempool min fee and w=
ill
>>>> be rejected, potentially breaking the acceptance expectations of the
>>>> package issuer ?
>>>>
>>>> Correct, if B+C is too low feerate to be accepted, we will reject it. =
I
>>>> prefer this because it is incentive compatible: A can be mined by itse=
lf,
>>>> so there's no reason to prefer A+B+C instead of A.
>>>> As another way of looking at this, consider the case where we do accep=
t
>>>> A+B+C and it sits at the "bottom" of our mempool. If our mempool reach=
es
>>>> capacity, we evict the lowest descendant feerate transactions, which a=
re
>>>> B+C in this case. This gives us the same resulting mempool, with A and=
 not
>>>> B+C.
>>>>
>>>>
>>>> > Further, I think the dedup should be done on wtxid, as you might hav=
e
>>>> multiple valid witnesses. Though with varying vsizes and as such offer=
ing
>>>> different feerates.
>>>>
>>>> I agree that variations of the same package with different witnesses i=
s
>>>> a case that must be handled. I consider witness replacement to be a pr=
oject
>>>> that can be done in parallel to package mempool acceptance because bei=
ng
>>>> able to accept packages does not worsen the problem of a
>>>> same-txid-different-witness "pinning" attack.
>>>>
>>>> If or when we have witness replacement, the logic is: if the individua=
l
>>>> transaction is enough to replace the mempool one, the replacement will
>>>> happen during the preceding individual transaction acceptance, and
>>>> deduplication logic will work. Otherwise, we will try to deduplicate b=
y
>>>> wtxid, see that we need a package witness replacement, and use the pac=
kage
>>>> feerate to evaluate whether this is economically rational.
>>>>
>>>> See the #22290 "handle package transactions already in mempool" commit=
 (
>>>> https://github.com/bitcoin/bitcoin/pull/22290/commits/fea75a2237b46cf7=
6145242fecad7e274bfcb5ff),
>>>> which handles the case of same-txid-different-witness by simply using =
the
>>>> transaction in the mempool for now, with TODOs for what I just describ=
ed.
>>>>
>>>>
>>>> > I'm not clearly understanding the accepted topologies. By "parent an=
d
>>>> child to share a parent", do you mean the set of transactions A, B, C,
>>>> where B is spending A and C is spending A and B would be correct ?
>>>>
>>>> Yes, that is what I meant. Yes, that would a valid package under these
>>>> rules.
>>>>
>>>> > If yes, is there a width-limit introduced or we fallback on
>>>> MAX_PACKAGE_COUNT=3D25 ?
>>>>
>>>> No, there is no limit on connectivity other than "child with all
>>>> unconfirmed parents." We will enforce MAX_PACKAGE_COUNT=3D25 and child=
's
>>>> in-mempool + in-package ancestor limits.
>>>>
>>>>
>>>> > Considering the current Core's mempool acceptance rules, I think CPF=
P
>>>> batching is unsafe for LN time-sensitive closure. A malicious tx-relay
>>>> jamming successful on one channel commitment transaction would contami=
ne
>>>> the remaining commitments sharing the same package.
>>>>
>>>> > E.g, you broadcast the package A+B+C+D+E where A,B,C,D are commitmen=
t
>>>> transactions and E a shared CPFP. If a malicious A' transaction has a
>>>> better feerate than A, the whole package acceptance will fail. Even if=
 A'
>>>> confirms in the following block,
>>>> the propagation and confirmation of B+C+D have been delayed. This coul=
d
>>>> carry on a loss of funds.
>>>>
>>>> Please note that A may replace A' even if A' has higher fees than A
>>>> individually, because the proposed package RBF utilizes the fees and s=
ize
>>>> of the entire package. This just requires E to pay enough fees, althou=
gh
>>>> this can be pretty high if there are also potential B' and C' competin=
g
>>>> commitment transactions that we don't know about.
>>>>
>>>>
>>>> > IMHO, I'm leaning towards deploying during a first phase
>>>> 1-parent/1-child. I think it's the most conservative step still improv=
ing
>>>> second-layer safety.
>>>>
>>>> So far, my understanding is that multi-parent-1-child is desired for
>>>> batched fee-bumping (
>>>> https://github.com/bitcoin/bitcoin/pull/22674#issuecomment-897951289)
>>>> and I've also seen your response which I have less context on (
>>>> https://github.com/bitcoin/bitcoin/pull/22674#issuecomment-900352202).
>>>> That being said, I am happy to create a new proposal for 1 parent + 1 =
child
>>>> (which would be slightly simpler) and plan for moving to
>>>> multi-parent-1-child later if that is preferred. I am very interested =
in
>>>> hearing feedback on that approach.
>>>>
>>>>
>>>> > If A+B is submitted to replace A', where A pays 0 sats, B pays 200
>>>> sats and A' pays 100 sats. If we apply the individual RBF on A, A+B
>>>> acceptance fails. For this reason I think the individual RBF should be
>>>> bypassed and only the package RBF apply ?
>>>>
>>>> I think there is a misunderstanding here - let me describe what I'm
>>>> proposing we'd do in this situation: we'll try individual submission f=
or A,
>>>> see that it fails due to "insufficient fees." Then, we'll try package
>>>> validation for A+B and use package RBF. If A+B pays enough, it can sti=
ll
>>>> replace A'. If A fails for a bad signature, we won't look at B or A+B.=
 Does
>>>> this meet your expectations?
>>>>
>>>>
>>>> > What problem are you trying to solve by the package feerate *after*
>>>> dedup rule ?
>>>> > My understanding is that an in-package transaction might be already
>>>> in the mempool. Therefore, to compute a correct RBF penalty replacemen=
t,
>>>> the vsize of this transaction could be discarded lowering the cost of
>>>> package RBF.
>>>>
>>>> I'm proposing that, when a transaction has already been submitted to
>>>> mempool, we would ignore both its fees and vsize when calculating pack=
age
>>>> feerate. In example G2, we shouldn't count M1 fees after its submissio=
n to
>>>> mempool, since M1's fees have already been used to pay for its individ=
ual
>>>> bandwidth, and it shouldn't be used again to pay for P2 and P3's bandw=
idth.
>>>> We also shouldn't count its vsize, since it has already been paid for.
>>>>
>>>>
>>>> > I think this is a footgunish API, as if a package issuer send the
>>>> multiple-parent-one-child package A,B,C,D where D is the child of A,B,=
C.
>>>> Then try to broadcast the higher-feerate C'+D' package, it should be
>>>> rejected. So it's breaking the naive broadcaster assumption that a
>>>> higher-feerate/higher-fee package always replaces ?
>>>>
>>>> Note that, if C' conflicts with C, it also conflicts with D, since D i=
s
>>>> a descendant of C and would thus need to be evicted along with it.
>>>> Implicitly, D' would not be in conflict with D.
>>>> More generally, this example is surprising to me because I didn't thin=
k
>>>> packages would be used to fee-bump replaceable transactions. Do we wan=
t the
>>>> child to be able to replace mempool transactions as well? This can be
>>>> implemented with a bit of additional logic.
>>>>
>>>> > I think this is unsafe for L2s if counterparties have malleability o=
f
>>>> the child transaction. They can block your package replacement by
>>>> opting-out from RBF signaling. IIRC, LN's "anchor output" presents suc=
h an
>>>> ability.
>>>>
>>>> I'm not sure what you mean? Let's say we have a package of parent A +
>>>> child B, where A is supposed to replace a mempool transaction A'. Are =
you
>>>> saying that counterparties are able to malleate the package child B, o=
r a
>>>> child of A'? If they can malleate a child of A', that shouldn't matter=
 as
>>>> long as A' is signaling replacement. This would be handled identically=
 with
>>>> full RBF and what Core currently implements.
>>>>
>>>> > I think this is an issue brought by the trimming during the dedup
>>>> phase. If we preserve the package integrity, only re-using the tx-leve=
l
>>>> checks results of already in-mempool transactions to gain in CPU time =
we
>>>> won't have this issue. Package childs can add unconfirmed inputs as lo=
ng as
>>>> they're in-package, the bip125 rule2 is only evaluated against parents=
 ?
>>>>
>>>> Sorry, I don't understand what you mean by "preserve the package
>>>> integrity?" Could you elaborate?
>>>>
>>>> > Let's say you have in-mempool A, B where A pays 10 sat/vb for 100
>>>> vbytes and B pays 10 sat/vb for 100 vbytes. You have the candidate
>>>> replacement D spending both A and C where D pays 15sat/vb for 100 vbyt=
es
>>>> and C pays 1 sat/vb for 1000 vbytes.
>>>>
>>>> > Package A + B ancestor score is 10 sat/vb.
>>>>
>>>> > D has a higher feerate/absolute fee than B.
>>>>
>>>> > Package A + C + D ancestor score is ~ 3 sat/vb ((A's 1000 sats + C's
>>>> 1000 sats + D's 1500 sats) / A's 100 vb + C's 1000 vb + D's 100 vb)
>>>>
>>>> I am in agreement with your calculations but unsure if we disagree on
>>>> the expected outcome. Yes, B has an ancestor score of 10sat/vb and D h=
as an
>>>> ancestor score of ~2.9sat/vb. Since D's ancestor score is lower than B=
's,
>>>> it fails the proposed package RBF Rule #2, so this package would be
>>>> rejected. Does this meet your expectations?
>>>>
>>>> Thank you for linking to projects that might be interested in package
>>>> relay :)
>>>>
>>>> Thanks,
>>>> Gloria
>>>>
>>>> On Mon, Sep 20, 2021 at 12:16 AM Antoine Riard <antoine.riard@gmail.co=
m>
>>>> wrote:
>>>>
>>>>> Hi Gloria,
>>>>>
>>>>> > A package may contain transactions that are already in the mempool.
>>>>> We
>>>>> > remove
>>>>> > ("deduplicate") those transactions from the package for the purpose=
s
>>>>> of
>>>>> > package
>>>>> > mempool acceptance. If a package is empty after deduplication, we d=
o
>>>>> > nothing.
>>>>>
>>>>> IIUC, you have a package A+B+C submitted for acceptance and A is
>>>>> already in your mempool. You trim out A from the package and then eva=
luate
>>>>> B+C.
>>>>>
>>>>> I think this might be an issue if A is the higher-fee element of the
>>>>> ABC package. B+C package fees might be under the mempool min fee and =
will
>>>>> be rejected, potentially breaking the acceptance expectations of the
>>>>> package issuer ?
>>>>>
>>>>> Further, I think the dedup should be done on wtxid, as you might have
>>>>> multiple valid witnesses. Though with varying vsizes and as such offe=
ring
>>>>> different feerates.
>>>>>
>>>>> E.g you're going to evaluate the package A+B and A' is already in you=
r
>>>>> mempool with a bigger valid witness. You trim A based on txid, then y=
ou
>>>>> evaluate A'+B, which fails the fee checks. However, evaluating A+B wo=
uld
>>>>> have been a success.
>>>>>
>>>>> AFAICT, the dedup rationale would be to save on CPU time/IO disk, to
>>>>> avoid repeated signatures verification and parent UTXOs fetches ? Can=
 we
>>>>> achieve the same goal by bypassing tx-level checks for already-in txn=
 while
>>>>> conserving the package integrity for package-level checks ?
>>>>>
>>>>> > Note that it's possible for the parents to be
>>>>> > indirect
>>>>> > descendants/ancestors of one another, or for parent and child to
>>>>> share a
>>>>> > parent,
>>>>> > so we cannot make any other topology assumptions.
>>>>>
>>>>> I'm not clearly understanding the accepted topologies. By "parent and
>>>>> child to share a parent", do you mean the set of transactions A, B, C=
,
>>>>> where B is spending A and C is spending A and B would be correct ?
>>>>>
>>>>> If yes, is there a width-limit introduced or we fallback on
>>>>> MAX_PACKAGE_COUNT=3D25 ?
>>>>>
>>>>> IIRC, one rationale to come with this topology limitation was to lowe=
r
>>>>> the DoS risks when potentially deploying p2p packages.
>>>>>
>>>>> Considering the current Core's mempool acceptance rules, I think CPFP
>>>>> batching is unsafe for LN time-sensitive closure. A malicious tx-rela=
y
>>>>> jamming successful on one channel commitment transaction would contam=
ine
>>>>> the remaining commitments sharing the same package.
>>>>>
>>>>> E.g, you broadcast the package A+B+C+D+E where A,B,C,D are commitment
>>>>> transactions and E a shared CPFP. If a malicious A' transaction has a
>>>>> better feerate than A, the whole package acceptance will fail. Even i=
f A'
>>>>> confirms in the following block,
>>>>> the propagation and confirmation of B+C+D have been delayed. This
>>>>> could carry on a loss of funds.
>>>>>
>>>>> That said, if you're broadcasting commitment transactions without
>>>>> time-sensitive HTLC outputs, I think the batching is effectively a fe=
e
>>>>> saving as you don't have to duplicate the CPFP.
>>>>>
>>>>> IMHO, I'm leaning towards deploying during a first phase
>>>>> 1-parent/1-child. I think it's the most conservative step still impro=
ving
>>>>> second-layer safety.
>>>>>
>>>>> > *Rationale*:  It would be incorrect to use the fees of transactions
>>>>> that are
>>>>> > already in the mempool, as we do not want a transaction's fees to b=
e
>>>>> > double-counted for both its individual RBF and package RBF.
>>>>>
>>>>> I'm unsure about the logical order of the checks proposed.
>>>>>
>>>>> If A+B is submitted to replace A', where A pays 0 sats, B pays 200
>>>>> sats and A' pays 100 sats. If we apply the individual RBF on A, A+B
>>>>> acceptance fails. For this reason I think the individual RBF should b=
e
>>>>> bypassed and only the package RBF apply ?
>>>>>
>>>>> Note this situation is plausible, with current LN design, your
>>>>> counterparty can have a commitment transaction with a better fee just=
 by
>>>>> selecting a higher `dust_limit_satoshis` than yours.
>>>>>
>>>>> > Examples F and G [14] show the same package, but P1 is submitted
>>>>> > individually before
>>>>> > the package in example G. In example F, we can see that the 300vB
>>>>> package
>>>>> > pays
>>>>> > an additional 200sat in fees, which is not enough to pay for its ow=
n
>>>>> > bandwidth
>>>>> > (BIP125#4). In example G, we can see that P1 pays enough to replace
>>>>> M1, but
>>>>> > using P1's fees again during package submission would make it look
>>>>> like a
>>>>> > 300sat
>>>>> > increase for a 200vB package. Even including its fees and size woul=
d
>>>>> not be
>>>>> > sufficient in this example, since the 300sat looks like enough for
>>>>> the 300vB
>>>>> > package. The calculcation after deduplication is 100sat increase fo=
r
>>>>> a
>>>>> > package
>>>>> > of size 200vB, which correctly fails BIP125#4. Assume all
>>>>> transactions have
>>>>> > a
>>>>> > size of 100vB.
>>>>>
>>>>> What problem are you trying to solve by the package feerate *after*
>>>>> dedup rule ?
>>>>>
>>>>> My understanding is that an in-package transaction might be already i=
n
>>>>> the mempool. Therefore, to compute a correct RBF penalty replacement,=
 the
>>>>> vsize of this transaction could be discarded lowering the cost of pac=
kage
>>>>> RBF.
>>>>>
>>>>> If we keep a "safe" dedup mechanism (see my point above), I think thi=
s
>>>>> discount is justified, as the validation cost of node operators is pa=
id for
>>>>> ?
>>>>>
>>>>> > The child cannot replace mempool transactions.
>>>>>
>>>>> Let's say you issue package A+B, then package C+B', where B' is a
>>>>> child of both A and C. This rule fails the acceptance of C+B' ?
>>>>>
>>>>> I think this is a footgunish API, as if a package issuer send the
>>>>> multiple-parent-one-child package A,B,C,D where D is the child of A,B=
,C.
>>>>> Then try to broadcast the higher-feerate C'+D' package, it should be
>>>>> rejected. So it's breaking the naive broadcaster assumption that a
>>>>> higher-feerate/higher-fee package always replaces ? And it might be u=
nsafe
>>>>> in protocols where states are symmetric. E.g a malicious counterparty
>>>>> broadcasts first S+A, then you honestly broadcast S+B, where B pays b=
etter
>>>>> fees.
>>>>>
>>>>> > All mempool transactions to be replaced must signal replaceability.
>>>>>
>>>>> I think this is unsafe for L2s if counterparties have malleability of
>>>>> the child transaction. They can block your package replacement by
>>>>> opting-out from RBF signaling. IIRC, LN's "anchor output" presents su=
ch an
>>>>> ability.
>>>>>
>>>>> I think it's better to either fix inherited signaling or move towards
>>>>> full-rbf.
>>>>>
>>>>> > if a package parent has already been submitted, it would
>>>>> > look
>>>>> >like the child is spending a "new" unconfirmed input.
>>>>>
>>>>> I think this is an issue brought by the trimming during the dedup
>>>>> phase. If we preserve the package integrity, only re-using the tx-lev=
el
>>>>> checks results of already in-mempool transactions to gain in CPU time=
 we
>>>>> won't have this issue. Package childs can add unconfirmed inputs as l=
ong as
>>>>> they're in-package, the bip125 rule2 is only evaluated against parent=
s ?
>>>>>
>>>>> > However, we still achieve the same goal of requiring the
>>>>> > replacement
>>>>> > transactions to have a ancestor score at least as high as the
>>>>> original
>>>>> > ones.
>>>>>
>>>>> I'm not sure if this holds...
>>>>>
>>>>> Let's say you have in-mempool A, B where A pays 10 sat/vb for 100
>>>>> vbytes and B pays 10 sat/vb for 100 vbytes. You have the candidate
>>>>> replacement D spending both A and C where D pays 15sat/vb for 100 vby=
tes
>>>>> and C pays 1 sat/vb for 1000 vbytes.
>>>>>
>>>>> Package A + B ancestor score is 10 sat/vb.
>>>>>
>>>>> D has a higher feerate/absolute fee than B.
>>>>>
>>>>> Package A + C + D ancestor score is ~ 3 sat/vb ((A's 1000 sats + C's
>>>>> 1000 sats + D's 1500 sats) /
>>>>> A's 100 vb + C's 1000 vb + D's 100 vb)
>>>>>
>>>>> Overall, this is a review through the lenses of LN requirements. I
>>>>> think other L2 protocols/applications
>>>>> could be candidates to using package accept/relay such as:
>>>>> * https://github.com/lightninglabs/pool
>>>>> * https://github.com/discreetlogcontracts/dlcspecs
>>>>> * https://github.com/bitcoin-teleport/teleport-transactions/
>>>>> * https://github.com/sapio-lang/sapio
>>>>> *
>>>>> https://github.com/commerceblock/mercury/blob/master/doc/statechains.=
md
>>>>> * https://github.com/revault/practical-revault
>>>>>
>>>>> Thanks for rolling forward the ball on this subject.
>>>>>
>>>>> Antoine
>>>>>
>>>>> Le jeu. 16 sept. 2021 =C3=A0 03:55, Gloria Zhao via bitcoin-dev <
>>>>> bitcoin-dev@lists.linuxfoundation.org> a =C3=A9crit :
>>>>>
>>>>>> Hi there,
>>>>>>
>>>>>> I'm writing to propose a set of mempool policy changes to enable
>>>>>> package
>>>>>> validation (in preparation for package relay) in Bitcoin Core. These
>>>>>> would not
>>>>>> be consensus or P2P protocol changes. However, since mempool policy
>>>>>> significantly affects transaction propagation, I believe this is
>>>>>> relevant for
>>>>>> the mailing list.
>>>>>>
>>>>>> My proposal enables packages consisting of multiple parents and 1
>>>>>> child. If you
>>>>>> develop software that relies on specific transaction relay
>>>>>> assumptions and/or
>>>>>> are interested in using package relay in the future, I'm very
>>>>>> interested to hear
>>>>>> your feedback on the utility or restrictiveness of these package
>>>>>> policies for
>>>>>> your use cases.
>>>>>>
>>>>>> A draft implementation of this proposal can be found in [Bitcoin Cor=
e
>>>>>> PR#22290][1].
>>>>>>
>>>>>> An illustrated version of this post can be found at
>>>>>> https://gist.github.com/glozow/dc4e9d5c5b14ade7cdfac40f43adb18a.
>>>>>> I have also linked the images below.
>>>>>>
>>>>>> ## Background
>>>>>>
>>>>>> Feel free to skip this section if you are already familiar with
>>>>>> mempool policy
>>>>>> and package relay terminology.
>>>>>>
>>>>>> ### Terminology Clarifications
>>>>>>
>>>>>> * Package =3D an ordered list of related transactions, representable=
 by
>>>>>> a Directed
>>>>>>   Acyclic Graph.
>>>>>> * Package Feerate =3D the total modified fees divided by the total
>>>>>> virtual size of
>>>>>>   all transactions in the package.
>>>>>>     - Modified fees =3D a transaction's base fees + fee delta applie=
d
>>>>>> by the user
>>>>>>       with `prioritisetransaction`. As such, we expect this to vary
>>>>>> across
>>>>>> mempools.
>>>>>>     - Virtual Size =3D the maximum of virtual sizes calculated using
>>>>>> [BIP141
>>>>>>       virtual size][2] and sigop weight. [Implemented here in Bitcoi=
n
>>>>>> Core][3].
>>>>>>     - Note that feerate is not necessarily based on the base fees an=
d
>>>>>> serialized
>>>>>>       size.
>>>>>>
>>>>>> * Fee-Bumping =3D user/wallet actions that take advantage of miner
>>>>>> incentives to
>>>>>>   boost a transaction's candidacy for inclusion in a block, includin=
g
>>>>>> Child Pays
>>>>>> for Parent (CPFP) and [BIP125][12] Replace-by-Fee (RBF). Our
>>>>>> intention in
>>>>>> mempool policy is to recognize when the new transaction is more
>>>>>> economical to
>>>>>> mine than the original one(s) but not open DoS vectors, so there are
>>>>>> some
>>>>>> limitations.
>>>>>>
>>>>>> ### Policy
>>>>>>
>>>>>> The purpose of the mempool is to store the best (to be most
>>>>>> incentive-compatible
>>>>>> with miners, highest feerate) candidates for inclusion in a block.
>>>>>> Miners use
>>>>>> the mempool to build block templates. The mempool is also useful as =
a
>>>>>> cache for
>>>>>> boosting block relay and validation performance, aiding transaction
>>>>>> relay, and
>>>>>> generating feerate estimations.
>>>>>>
>>>>>> Ideally, all consensus-valid transactions paying reasonable fees
>>>>>> should make it
>>>>>> to miners through normal transaction relay, without any special
>>>>>> connectivity or
>>>>>> relationships with miners. On the other hand, nodes do not have
>>>>>> unlimited
>>>>>> resources, and a P2P network designed to let any honest node
>>>>>> broadcast their
>>>>>> transactions also exposes the transaction validation engine to DoS
>>>>>> attacks from
>>>>>> malicious peers.
>>>>>>
>>>>>> As such, for unconfirmed transactions we are considering for our
>>>>>> mempool, we
>>>>>> apply a set of validation rules in addition to consensus, primarily
>>>>>> to protect
>>>>>> us from resource exhaustion and aid our efforts to keep the highest
>>>>>> fee
>>>>>> transactions. We call this mempool _policy_: a set of (configurable,
>>>>>> node-specific) rules that transactions must abide by in order to be
>>>>>> accepted
>>>>>> into our mempool. Transaction "Standardness" rules and mempool
>>>>>> restrictions such
>>>>>> as "too-long-mempool-chain" are both examples of policy.
>>>>>>
>>>>>> ### Package Relay and Package Mempool Accept
>>>>>>
>>>>>> In transaction relay, we currently consider transactions one at a
>>>>>> time for
>>>>>> submission to the mempool. This creates a limitation in the node's
>>>>>> ability to
>>>>>> determine which transactions have the highest feerates, since we
>>>>>> cannot take
>>>>>> into account descendants (i.e. cannot use CPFP) until all the
>>>>>> transactions are
>>>>>> in the mempool. Similarly, we cannot use a transaction's descendants
>>>>>> when
>>>>>> considering it for RBF. When an individual transaction does not meet
>>>>>> the mempool
>>>>>> minimum feerate and the user isn't able to create a replacement
>>>>>> transaction
>>>>>> directly, it will not be accepted by mempools.
>>>>>>
>>>>>> This limitation presents a security issue for applications and users
>>>>>> relying on
>>>>>> time-sensitive transactions. For example, Lightning and other
>>>>>> protocols create
>>>>>> UTXOs with multiple spending paths, where one counterparty's spendin=
g
>>>>>> path opens
>>>>>> up after a timelock, and users are protected from cheating scenarios
>>>>>> as long as
>>>>>> they redeem on-chain in time. A key security assumption is that all
>>>>>> parties'
>>>>>> transactions will propagate and confirm in a timely manner. This
>>>>>> assumption can
>>>>>> be broken if fee-bumping does not work as intended.
>>>>>>
>>>>>> The end goal for Package Relay is to consider multiple transactions
>>>>>> at the same
>>>>>> time, e.g. a transaction with its high-fee child. This may help us
>>>>>> better
>>>>>> determine whether transactions should be accepted to our mempool,
>>>>>> especially if
>>>>>> they don't meet fee requirements individually or are better RBF
>>>>>> candidates as a
>>>>>> package. A combination of changes to mempool validation logic,
>>>>>> policy, and
>>>>>> transaction relay allows us to better propagate the transactions wit=
h
>>>>>> the
>>>>>> highest package feerates to miners, and makes fee-bumping tools more
>>>>>> powerful
>>>>>> for users.
>>>>>>
>>>>>> The "relay" part of Package Relay suggests P2P messaging changes, bu=
t
>>>>>> a large
>>>>>> part of the changes are in the mempool's package validation logic. W=
e
>>>>>> call this
>>>>>> *Package Mempool Accept*.
>>>>>>
>>>>>> ### Previous Work
>>>>>>
>>>>>> * Given that mempool validation is DoS-sensitive and complex, it
>>>>>> would be
>>>>>>   dangerous to haphazardly tack on package validation logic. Many
>>>>>> efforts have
>>>>>> been made to make mempool validation less opaque (see [#16400][4],
>>>>>> [#21062][5],
>>>>>> [#22675][6], [#22796][7]).
>>>>>> * [#20833][8] Added basic capabilities for package validation, test
>>>>>> accepts only
>>>>>>   (no submission to mempool).
>>>>>> * [#21800][9] Implemented package ancestor/descendant limit checks
>>>>>> for arbitrary
>>>>>>   packages. Still test accepts only.
>>>>>> * Previous package relay proposals (see [#16401][10], [#19621][11]).
>>>>>>
>>>>>> ### Existing Package Rules
>>>>>>
>>>>>> These are in master as introduced in [#20833][8] and [#21800][9].
>>>>>> I'll consider
>>>>>> them as "given" in the rest of this document, though they can be
>>>>>> changed, since
>>>>>> package validation is test-accept only right now.
>>>>>>
>>>>>> 1. A package cannot exceed `MAX_PACKAGE_COUNT=3D25` count and
>>>>>> `MAX_PACKAGE_SIZE=3D101KvB` total size [8]
>>>>>>
>>>>>>    *Rationale*: This is already enforced as mempool
>>>>>> ancestor/descendant limits.
>>>>>> Presumably, transactions in a package are all related, so exceeding
>>>>>> this limit
>>>>>> would mean that the package can either be split up or it wouldn't
>>>>>> pass this
>>>>>> mempool policy.
>>>>>>
>>>>>> 2. Packages must be topologically sorted: if any dependencies exist
>>>>>> between
>>>>>> transactions, parents must appear somewhere before children. [8]
>>>>>>
>>>>>> 3. A package cannot have conflicting transactions, i.e. none of them
>>>>>> can spend
>>>>>> the same inputs. This also means there cannot be duplicate
>>>>>> transactions. [8]
>>>>>>
>>>>>> 4. When packages are evaluated against ancestor/descendant limits in
>>>>>> a test
>>>>>> accept, the union of all of their descendants and ancestors is
>>>>>> considered. This
>>>>>> is essentially a "worst case" heuristic where every transaction in
>>>>>> the package
>>>>>> is treated as each other's ancestor and descendant. [8]
>>>>>> Packages for which ancestor/descendant limits are accurately capture=
d
>>>>>> by this
>>>>>> heuristic: [19]
>>>>>>
>>>>>> There are also limitations such as the fact that CPFP carve out is
>>>>>> not applied
>>>>>> to package transactions. #20833 also disables RBF in package
>>>>>> validation; this
>>>>>> proposal overrides that to allow packages to use RBF.
>>>>>>
>>>>>> ## Proposed Changes
>>>>>>
>>>>>> The next step in the Package Mempool Accept project is to implement
>>>>>> submission
>>>>>> to mempool, initially through RPC only. This allows us to test the
>>>>>> submission
>>>>>> logic before exposing it on P2P.
>>>>>>
>>>>>> ### Summary
>>>>>>
>>>>>> - Packages may contain already-in-mempool transactions.
>>>>>> - Packages are 2 generations, Multi-Parent-1-Child.
>>>>>> - Fee-related checks use the package feerate. This means that wallet=
s
>>>>>> can
>>>>>> create a package that utilizes CPFP.
>>>>>> - Parents are allowed to RBF mempool transactions with a set of rule=
s
>>>>>> similar
>>>>>>   to BIP125. This enables a combination of CPFP and RBF, where a
>>>>>> transaction's descendant fees pay for replacing mempool conflicts.
>>>>>>
>>>>>> There is a draft implementation in [#22290][1]. It is WIP, but
>>>>>> feedback is
>>>>>> always welcome.
>>>>>>
>>>>>> ### Details
>>>>>>
>>>>>> #### Packages May Contain Already-in-Mempool Transactions
>>>>>>
>>>>>> A package may contain transactions that are already in the mempool.
>>>>>> We remove
>>>>>> ("deduplicate") those transactions from the package for the purposes
>>>>>> of package
>>>>>> mempool acceptance. If a package is empty after deduplication, we do
>>>>>> nothing.
>>>>>>
>>>>>> *Rationale*: Mempools vary across the network. It's possible for a
>>>>>> parent to be
>>>>>> accepted to the mempool of a peer on its own due to differences in
>>>>>> policy and
>>>>>> fee market fluctuations. We should not reject or penalize the entire
>>>>>> package for
>>>>>> an individual transaction as that could be a censorship vector.
>>>>>>
>>>>>> #### Packages Are Multi-Parent-1-Child
>>>>>>
>>>>>> Only packages of a specific topology are permitted. Namely, a packag=
e
>>>>>> is exactly
>>>>>> 1 child with all of its unconfirmed parents. After deduplication, th=
e
>>>>>> package
>>>>>> may be exactly the same, empty, 1 child, 1 child with just some of i=
ts
>>>>>> unconfirmed parents, etc. Note that it's possible for the parents to
>>>>>> be indirect
>>>>>> descendants/ancestors of one another, or for parent and child to
>>>>>> share a parent,
>>>>>> so we cannot make any other topology assumptions.
>>>>>>
>>>>>> *Rationale*: This allows for fee-bumping by CPFP. Allowing multiple
>>>>>> parents
>>>>>> makes it possible to fee-bump a batch of transactions. Restricting
>>>>>> packages to a
>>>>>> defined topology is also easier to reason about and simplifies the
>>>>>> validation
>>>>>> logic greatly. Multi-parent-1-child allows us to think of the packag=
e
>>>>>> as one big
>>>>>> transaction, where:
>>>>>>
>>>>>> - Inputs =3D all the inputs of parents + inputs of the child that co=
me
>>>>>> from
>>>>>>   confirmed UTXOs
>>>>>> - Outputs =3D all the outputs of the child + all outputs of the pare=
nts
>>>>>> that
>>>>>>   aren't spent by other transactions in the package
>>>>>>
>>>>>> Examples of packages that follow this rule (variations of example A
>>>>>> show some
>>>>>> possibilities after deduplication): ![image][15]
>>>>>>
>>>>>> #### Fee-Related Checks Use Package Feerate
>>>>>>
>>>>>> Package Feerate =3D the total modified fees divided by the total
>>>>>> virtual size of
>>>>>> all transactions in the package.
>>>>>>
>>>>>> To meet the two feerate requirements of a mempool, i.e., the
>>>>>> pre-configured
>>>>>> minimum relay feerate (`minRelayTxFee`) and dynamic mempool minimum
>>>>>> feerate, the
>>>>>> total package feerate is used instead of the individual feerate. The
>>>>>> individual
>>>>>> transactions are allowed to be below feerate requirements if the
>>>>>> package meets
>>>>>> the feerate requirements. For example, the parent(s) in the package
>>>>>> can have 0
>>>>>> fees but be paid for by the child.
>>>>>>
>>>>>> *Rationale*: This can be thought of as "CPFP within a package,"
>>>>>> solving the
>>>>>> issue of a parent not meeting minimum fees on its own. This allows L=
2
>>>>>> applications to adjust their fees at broadcast time instead of
>>>>>> overshooting or
>>>>>> risking getting stuck/pinned.
>>>>>>
>>>>>> We use the package feerate of the package *after deduplication*.
>>>>>>
>>>>>> *Rationale*:  It would be incorrect to use the fees of transactions
>>>>>> that are
>>>>>> already in the mempool, as we do not want a transaction's fees to be
>>>>>> double-counted for both its individual RBF and package RBF.
>>>>>>
>>>>>> Examples F and G [14] show the same package, but P1 is submitted
>>>>>> individually before
>>>>>> the package in example G. In example F, we can see that the 300vB
>>>>>> package pays
>>>>>> an additional 200sat in fees, which is not enough to pay for its own
>>>>>> bandwidth
>>>>>> (BIP125#4). In example G, we can see that P1 pays enough to replace
>>>>>> M1, but
>>>>>> using P1's fees again during package submission would make it look
>>>>>> like a 300sat
>>>>>> increase for a 200vB package. Even including its fees and size would
>>>>>> not be
>>>>>> sufficient in this example, since the 300sat looks like enough for
>>>>>> the 300vB
>>>>>> package. The calculcation after deduplication is 100sat increase for
>>>>>> a package
>>>>>> of size 200vB, which correctly fails BIP125#4. Assume all
>>>>>> transactions have a
>>>>>> size of 100vB.
>>>>>>
>>>>>> #### Package RBF
>>>>>>
>>>>>> If a package meets feerate requirements as a package, the parents in
>>>>>> the
>>>>>> transaction are allowed to replace-by-fee mempool transactions. The
>>>>>> child cannot
>>>>>> replace mempool transactions. Multiple transactions can replace the
>>>>>> same
>>>>>> transaction, but in order to be valid, none of the transactions can
>>>>>> try to
>>>>>> replace an ancestor of another transaction in the same package (whic=
h
>>>>>> would thus
>>>>>> make its inputs unavailable).
>>>>>>
>>>>>> *Rationale*: Even if we are using package feerate, a package will no=
t
>>>>>> propagate
>>>>>> as intended if RBF still requires each individual transaction to mee=
t
>>>>>> the
>>>>>> feerate requirements.
>>>>>>
>>>>>> We use a set of rules slightly modified from BIP125 as follows:
>>>>>>
>>>>>> ##### Signaling (Rule #1)
>>>>>>
>>>>>> All mempool transactions to be replaced must signal replaceability.
>>>>>>
>>>>>> *Rationale*: Package RBF signaling logic should be the same for
>>>>>> package RBF and
>>>>>> single transaction acceptance. This would be updated if single
>>>>>> transaction
>>>>>> validation moves to full RBF.
>>>>>>
>>>>>> ##### New Unconfirmed Inputs (Rule #2)
>>>>>>
>>>>>> A package may include new unconfirmed inputs, but the ancestor
>>>>>> feerate of the
>>>>>> child must be at least as high as the ancestor feerates of every
>>>>>> transaction
>>>>>> being replaced. This is contrary to BIP125#2, which states "The
>>>>>> replacement
>>>>>> transaction may only include an unconfirmed input if that input was
>>>>>> included in
>>>>>> one of the original transactions. (An unconfirmed input spends an
>>>>>> output from a
>>>>>> currently-unconfirmed transaction.)"
>>>>>>
>>>>>> *Rationale*: The purpose of BIP125#2 is to ensure that the replaceme=
nt
>>>>>> transaction has a higher ancestor score than the original
>>>>>> transaction(s) (see
>>>>>> [comment][13]). Example H [16] shows how adding a new unconfirmed
>>>>>> input can lower the
>>>>>> ancestor score of the replacement transaction. P1 is trying to
>>>>>> replace M1, and
>>>>>> spends an unconfirmed output of M2. P1 pays 800sat, M1 pays 600sat,
>>>>>> and M2 pays
>>>>>> 100sat. Assume all transactions have a size of 100vB. While, in
>>>>>> isolation, P1
>>>>>> looks like a better mining candidate than M1, it must be mined with
>>>>>> M2, so its
>>>>>> ancestor feerate is actually 4.5sat/vB.  This is lower than M1's
>>>>>> ancestor
>>>>>> feerate, which is 6sat/vB.
>>>>>>
>>>>>> In package RBF, the rule analogous to BIP125#2 would be "none of the
>>>>>> transactions in the package can spend new unconfirmed inputs."
>>>>>> Example J [17] shows
>>>>>> why, if any of the package transactions have ancestors, package
>>>>>> feerate is no
>>>>>> longer accurate. Even though M2 and M3 are not ancestors of P1 (whic=
h
>>>>>> is the
>>>>>> replacement transaction in an RBF), we're actually interested in the
>>>>>> entire
>>>>>> package. A miner should mine M1 which is 5sat/vB instead of M2, M3,
>>>>>> P1, P2, and
>>>>>> P3, which is only 4sat/vB. The Package RBF rule cannot be loosened t=
o
>>>>>> only allow
>>>>>> the child to have new unconfirmed inputs, either, because it can
>>>>>> still cause us
>>>>>> to overestimate the package's ancestor score.
>>>>>>
>>>>>> However, enforcing a rule analogous to BIP125#2 would not only make
>>>>>> Package RBF
>>>>>> less useful, but would also break Package RBF for packages with
>>>>>> parents already
>>>>>> in the mempool: if a package parent has already been submitted, it
>>>>>> would look
>>>>>> like the child is spending a "new" unconfirmed input. In example K
>>>>>> [18], we're
>>>>>> looking to replace M1 with the entire package including P1, P2, and
>>>>>> P3. We must
>>>>>> consider the case where one of the parents is already in the mempool
>>>>>> (in this
>>>>>> case, P2), which means we must allow P3 to have new unconfirmed
>>>>>> inputs. However,
>>>>>> M2 lowers the ancestor score of P3 to 4.3sat/vB, so we should not
>>>>>> replace M1
>>>>>> with this package.
>>>>>>
>>>>>> Thus, the package RBF rule regarding new unconfirmed inputs is less
>>>>>> strict than
>>>>>> BIP125#2. However, we still achieve the same goal of requiring the
>>>>>> replacement
>>>>>> transactions to have a ancestor score at least as high as the
>>>>>> original ones. As
>>>>>> a result, the entire package is required to be a higher feerate
>>>>>> mining candidate
>>>>>> than each of the replaced transactions.
>>>>>>
>>>>>> Another note: the [comment][13] above the BIP125#2 code in the
>>>>>> original RBF
>>>>>> implementation suggests that the rule was intended to be temporary.
>>>>>>
>>>>>> ##### Absolute Fee (Rule #3)
>>>>>>
>>>>>> The package must increase the absolute fee of the mempool, i.e. the
>>>>>> total fees
>>>>>> of the package must be higher than the absolute fees of the mempool
>>>>>> transactions
>>>>>> it replaces. Combined with the CPFP rule above, this differs from
>>>>>> BIP125 Rule #3
>>>>>> - an individual transaction in the package may have lower fees than
>>>>>> the
>>>>>>   transaction(s) it is replacing. In fact, it may have 0 fees, and
>>>>>> the child
>>>>>> pays for RBF.
>>>>>>
>>>>>> ##### Feerate (Rule #4)
>>>>>>
>>>>>> The package must pay for its own bandwidth; the package feerate must
>>>>>> be higher
>>>>>> than the replaced transactions by at least minimum relay feerate
>>>>>> (`incrementalRelayFee`). Combined with the CPFP rule above, this
>>>>>> differs from
>>>>>> BIP125 Rule #4 - an individual transaction in the package can have a
>>>>>> lower
>>>>>> feerate than the transaction(s) it is replacing. In fact, it may hav=
e
>>>>>> 0 fees,
>>>>>> and the child pays for RBF.
>>>>>>
>>>>>> ##### Total Number of Replaced Transactions (Rule #5)
>>>>>>
>>>>>> The package cannot replace more than 100 mempool transactions. This
>>>>>> is identical
>>>>>> to BIP125 Rule #5.
>>>>>>
>>>>>> ### Expected FAQs
>>>>>>
>>>>>> 1. Is it possible for only some of the package to make it into the
>>>>>> mempool?
>>>>>>
>>>>>>    Yes, it is. However, since we evict transactions from the mempool
>>>>>> by
>>>>>> descendant score and the package child is supposed to be sponsoring
>>>>>> the fees of
>>>>>> its parents, the most common scenario would be all-or-nothing. This =
is
>>>>>> incentive-compatible. In fact, to be conservative, package validatio=
n
>>>>>> should
>>>>>> begin by trying to submit all of the transactions individually, and
>>>>>> only use the
>>>>>> package mempool acceptance logic if the parents fail due to low
>>>>>> feerate.
>>>>>>
>>>>>> 2. Should we allow packages to contain already-confirmed transaction=
s?
>>>>>>
>>>>>>     No, for practical reasons. In mempool validation, we actually
>>>>>> aren't able to
>>>>>> tell with 100% confidence if we are looking at a transaction that ha=
s
>>>>>> already
>>>>>> confirmed, because we look up inputs using a UTXO set. If we have
>>>>>> historical
>>>>>> block data, it's possible to look for it, but this is inefficient,
>>>>>> not always
>>>>>> possible for pruning nodes, and unnecessary because we're not going
>>>>>> to do
>>>>>> anything with the transaction anyway. As such, we already have the
>>>>>> expectation
>>>>>> that transaction relay is somewhat "stateful" i.e. nobody should be
>>>>>> relaying
>>>>>> transactions that have already been confirmed. Similarly, we
>>>>>> shouldn't be
>>>>>> relaying packages that contain already-confirmed transactions.
>>>>>>
>>>>>> [1]: https://github.com/bitcoin/bitcoin/pull/22290
>>>>>> [2]:
>>>>>> https://github.com/bitcoin/bips/blob/1f0b563738199ca60d32b4ba779797f=
c97d040fe/bip-0141.mediawiki#transaction-size-calculations
>>>>>> [3]:
>>>>>> https://github.com/bitcoin/bitcoin/blob/94f83534e4b771944af7d9ed0f40=
746f392eb75e/src/policy/policy.cpp#L282
>>>>>> [4]: https://github.com/bitcoin/bitcoin/pull/16400
>>>>>> [5]: https://github.com/bitcoin/bitcoin/pull/21062
>>>>>> [6]: https://github.com/bitcoin/bitcoin/pull/22675
>>>>>> [7]: https://github.com/bitcoin/bitcoin/pull/22796
>>>>>> [8]: https://github.com/bitcoin/bitcoin/pull/20833
>>>>>> [9]: https://github.com/bitcoin/bitcoin/pull/21800
>>>>>> [10]: https://github.com/bitcoin/bitcoin/pull/16401
>>>>>> [11]: https://github.com/bitcoin/bitcoin/pull/19621
>>>>>> [12]: https://github.com/bitcoin/bips/blob/master/bip-0125.mediawiki
>>>>>> [13]:
>>>>>> https://github.com/bitcoin/bitcoin/pull/6871/files#diff-34d21af3c614=
ea3cee120df276c9c4ae95053830d7f1d3deaf009a4625409ad2R1101-R1104
>>>>>> [14]:
>>>>>> https://user-images.githubusercontent.com/25183001/133567078-075a971=
c-0619-4339-9168-b41fd2b90c28.png
>>>>>> [15]:
>>>>>> https://user-images.githubusercontent.com/25183001/132856734-fc17da7=
5-f875-44bb-b954-cb7a1725cc0d.png
>>>>>> [16]:
>>>>>> https://user-images.githubusercontent.com/25183001/133567347-a3e2e4a=
8-ae9c-49f8-abb9-81e8e0aba224.png
>>>>>> [17]:
>>>>>> https://user-images.githubusercontent.com/25183001/133567370-21566d0=
e-36c8-4831-b1a8-706634540af3.png
>>>>>> [18]:
>>>>>> https://user-images.githubusercontent.com/25183001/133567444-bfff114=
2-439f-4547-800a-2ba2b0242bcb.png
>>>>>> [19]:
>>>>>> https://user-images.githubusercontent.com/25183001/133456219-0bb447c=
b-dcb4-4a31-b9c1-7d86205b68bc.png
>>>>>> [20]:
>>>>>> https://user-images.githubusercontent.com/25183001/132857787-7b7c6f5=
6-af96-44c8-8d78-983719888c19.png
>>>>>> _______________________________________________
>>>>>> bitcoin-dev mailing list
>>>>>> bitcoin-dev@lists.linuxfoundation.org
>>>>>> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>>>>>>
>>>>> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>

--000000000000a309eb05ccf4dc02
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px =
0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">I think =
we could restrain package acceptance to only confirmed inputs for now and r=
evisit later this point ? For LN-anchor, you can assume that the fee-bumpin=
g UTXO feeding the CPFP is already<br>confirmed. Or are there currently-dep=
loyed use-cases which would benefit from your proposed Rule #2 ?<br></block=
quote><div><br></div><div>I think constraining package acceptance to only c=
onfirmed inputs</div><div>is very limiting and quite dangerous for L2 proto=
cols.</div><div><br></div><div>In the case of LN, an attacker can game this=
 and heavily restrict</div><div>your RBF attempts if you&#39;re only allowe=
d to use confirmed inputs</div><div>and have many channels (and a limited n=
umber of confirmed inputs).</div><div>Otherwise you&#39;ll need node operat=
ors to pre-emptively split their</div><div>utxos into many small utxos just=
 for fee bumping, which is inefficient...</div><div><br></div><div>Bastien<=
/div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_a=
ttr">Le=C2=A0lun. 27 sept. 2021 =C3=A0=C2=A000:27, Antoine Riard via bitcoi=
n-dev &lt;<a href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org">bitcoin-=
dev@lists.linuxfoundation.org</a>&gt; a =C3=A9crit=C2=A0:<br></div><blockqu=
ote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px=
 solid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr">Hi Gloria,<br><b=
r>Thanks for your answers,<br><br>&gt; In summary, it seems that the decisi=
ons that might still need<br>&gt; attention/input from devs on this mailing=
 list are:<br>&gt; 1. Whether we should start with multiple-parent-1-child =
or 1-parent-1-child.<br>&gt; 2. Whether it&#39;s ok to require that the chi=
ld not have conflicts with<br>&gt; mempool transactions.<br><br>Yes 1) it w=
ould be good to have inputs of more potential users of package acceptance .=
 And 2) I think it&#39;s more a matter of clearer wording of the proposal.<=
br><br>However, see my final point on the relaxation around &quot;unconfirm=
ed inputs&quot; which might in fact alter our current block construction st=
rategy.<br><br>&gt; Right, the fact that we essentially always choose the f=
irst-seen witness is<br>&gt; an unfortunate limitation that exists already.=
 Adding package mempool<br>&gt; accept doesn&#39;t worsen this, but the pro=
cedure in the future is to replace<br>&gt; the witness when it makes sense =
economically. We can also add logic to<br>&gt; allow package feerate to pay=
 for witness replacements as well. This is<br>&gt; pretty far into the futu=
re, though.<br><br>Yes I agree package mempool doesn&#39;t worsen this. And=
 it&#39;s not an issue for current LN as you can&#39;t significantly inflat=
e a spending witness for the 2-of-2 funding output.<br>However, it might be=
 an issue for multi-party protocol where the spending script has alternativ=
e branches with asymmetric valid witness weights. Taproot should ease that =
kind of script so hopefully we would deploy wtxid-replacement not too far i=
n the future.<br><br>&gt; I could be misunderstanding, but an attacker woul=
dn&#39;t be able to<br>&gt; batch-attack like this. Alice&#39;s package onl=
y conflicts with A&#39; + D&#39;, not A&#39;<br>&gt; + B&#39; + C&#39; + D&=
#39;. She only needs to pay for evicting 2 transactions.<br><br>Yeah I can =
be clearer, I think you have 2 pinning attacks scenarios to consider.<br><b=
r>In LN, if you&#39;re trying to confirm a commitment transaction to time-o=
ut or claim on-chain a HTLC and the timelock is near-expiration, you should=
 be ready to pay in commitment+2nd-stage HTLC transaction fees as much as t=
he value offered by the HTLC.<br><br>Following this security assumption, an=
 attacker can exploit it by targeting together commitment transactions from=
 different channels by blocking them under a high-fee child, of which the f=
ee value<br>is equal to the top-value HTLC + 1. Victims&#39;s fee-bumping l=
ogics won&#39;t overbid as it&#39;s not worthy to offer fees beyond their c=
ompeted HTLCs. Apart from observing mempools state, victims can&#39;t learn=
 they&#39;re targeted by the same attacker.<br><br>To draw from the aforeme=
ntioned topology, Mallory broadcasts A&#39; + B&#39; + C&#39; + D&#39;, whe=
re A&#39; conflicts with Alice&#39;s P1, B&#39; conflicts with Bob&#39;s P2=
, C&#39; conflicts with Caroll&#39;s P3. Let&#39;s assume P1 is confirming =
the top-value HTLC of the set. If D&#39; fees is higher than P1 + 1, it won=
&#39;t be rational for Alice or Bob or Caroll to keep offering competing fe=
erates. Mallory will be at loss on stealing P1, as she has paid more in fee=
s but will realize a gain on P2+P3.<br><br>In this model, Alice is allowed =
to evict those 2 transactions (A&#39; + D&#39;) but as she is economically-=
bounded she won&#39;t succeed.<br><br>Mallory is maliciously exploiting RBF=
 rule 3 on absolute fee. I think this 1st pinning scenario is correct and &=
quot;lucractive&quot; when you sum the global gain/loss.<br><br>There is a =
2nd attack scenario where A + B + C + D, where D is the child of A,B,C. All=
 those transactions are honestly issued by Alice. Once A + B + C + D are pr=
opagated in network mempools, Mallory is able to replace A + D with =C2=A0A=
&#39; + D&#39; where D&#39; is paying a higher fee. This package A&#39; + D=
&#39; will confirm soon if D feerate was compelling but Mallory succeeds in=
 delaying the confirmation<br>of B + C for one or more blocks. As B + C are=
 pre-signed commitments with a low-fee rate they won&#39;t confirm without =
Alice issuing a new child E. Mallory can repeat the same trick by broadcast=
ing<br>B&#39; + E&#39; and delay again the confirmation of C.<br><br>If the=
 remaining package pending HTLC has a higher-value than all the malicious f=
ees over-bid, Mallory should realize a gain. With this 2nd pinning attack, =
the malicious entity buys confirmation delay of your packaged-together comm=
itments.<br><br>Assuming those attacks are correct, I&#39;m leaning towards=
 being conservative with the LDK broadcast backend. Though once again, othe=
r L2 devs have likely other use-cases and opinions :)<br><br>&gt; =C2=A0B&#=
39; only needs to pay for itself in this case.<br><br>Yes I think it&#39;s =
a nice discount when UTXO is single-owned. In the context of shared-owned U=
TXO (e.g LN), you might not if there is an in-mempool package already spend=
ing the UTXO and have to assume the worst-case scenario. I.e have B&#39; co=
mmitting enough fee to pay for A&#39; replacement bandwidth. I think we can=
&#39;t do that much for this case...<br><br>&gt; If a package meets feerate=
 requirements as a<br>package, the parents in the transaction are allowed t=
o replace-by-fee<br>mempool transactions. The child cannot replace mempool =
transactions.&quot;<br><br>I agree with the Mallory-vs-Alice case. Though i=
f Alice broadcasts A+B&#39; to replace A+B because the first broadcast isn&=
#39;t satisfying anymore due to mempool spikes ? Assuming B&#39; fees is en=
ough, I think that case as child B&#39; replacing in-mempool transaction B.=
 Which I understand going against=C2=A0 &quot;The child cannot replace memp=
ool transactions&quot;.<br><br>Maybe wording could be a bit clearer ?<br><b=
r>&gt; While it would be nice to have full RBF, malleability of the child w=
on&#39;t<br>&gt; block RBF here. If we&#39;re trying to replace A&#39;, we =
only require that A&#39;<br>&gt; signals replaceability, and don&#39;t mind=
 if its child doesn&#39;t.<br><br>Yes, it sounds good.<br><br>&gt; Yes, A+C=
+D pays 2500sat more in fees, but it is also 1000vB larger. A miner<br>&gt;=
 should prefer to utilize their block space more effectively.<br><br>If you=
r mempool is empty and only composed of A+C+D or A+B, I think taking A+C+D =
is the most efficient block construction you can come up with as a miner ?<=
br><br>&gt; No, because we don&#39;t use that model.<br><br>Can you describ=
e what miner model we are using ? Like the block construction strategy impl=
emented by `addPackagesTxs` or also encompassing our current mempool accept=
ance policy, which I think rely on absolute fee over ancestor score in case=
 of replacement ?<br><br>I think this point is worthy to discuss as otherwi=
se we might downgrade the efficiency of our current block construction stra=
tegy in periods of near-empty mempools. A knowledge which could be discreet=
ly leveraged by a miner to gain an advantage on the rest of the mining ecos=
ystem.<br><br>Note, I think we *might* have to go in this direction if we w=
ant to replace replace-by-fee by replace-by-feerate or replace-by-ancestor =
and solve in-depth pinning attacks. Though if we do so, <br>IMO we would ne=
ed more thoughts.<br><br>I think we could restrain package acceptance to on=
ly confirmed inputs for now and revisit later this point ? For LN-anchor, y=
ou can assume that the fee-bumping UTXO feeding the CPFP is already<br>conf=
irmed. Or are there currently-deployed use-cases which would benefit from y=
our proposed Rule #2 ?<br><br>Antoine<br></div><br><div class=3D"gmail_quot=
e"><div dir=3D"ltr" class=3D"gmail_attr">Le=C2=A0jeu. 23 sept. 2021 =C3=A0=
=C2=A011:36, Gloria Zhao &lt;<a href=3D"mailto:gloriajzhao@gmail.com" targe=
t=3D"_blank">gloriajzhao@gmail.com</a>&gt; a =C3=A9crit=C2=A0:<br></div><bl=
ockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-lef=
t:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr">Hi Antoine,=
<br><br><div>Thanks as always for your input. I&#39;m glad we agree on so m=
uch!</div><div><br></div><div>In summary, it seems that the decisions that =
might still need attention/input from devs on this mailing list are:</div>1=
. Whether we should start with multiple-parent-1-child or 1-parent-1-child.=
<br>2. Whether it&#39;s ok to require that the child not have conflicts wit=
h mempool transactions.<br><br><div>Responding to your comments...</div><di=
v><br></div>&gt; IIUC, you have package A+B, during the dedup phase early i=
n `AcceptMultipleTransactions` if you observe same-txid-different-wtixd A&#=
39; and A&#39; is higher feerate than A, you trim A and replace by A&#39; ?=
<br><br>&gt; I think this approach is safe, the one who appears unsafe to m=
e is when A&#39; has a _lower_ feerate, even if A&#39; is already accepted =
by our mempool ? In that case iirc that would be a pinning.<br><br>Right, t=
he fact that we essentially always choose the first-seen witness is an unfo=
rtunate limitation that exists already. Adding package mempool accept doesn=
&#39;t worsen this, but the procedure in the future is to replace the witne=
ss when it makes sense economically. We can also add logic to allow package=
 feerate to pay for witness replacements as well. This is pretty far into t=
he future, though.<br><br>&gt; It sounds uneconomical for an attacker but I=
 think it&#39;s not when you consider than you can &quot;batch&quot; attack=
 against multiple honest counterparties. E.g, Mallory broadcast A&#39; + B&=
#39; + C&#39; + D&#39; where A&#39; conflicts with Alice&#39;s honest packa=
ge P1, B&#39; conflicts with Bob&#39;s honest package P2, C&#39; conflicts =
with Caroll&#39;s honest package P3. And D&#39; is a high-fee child of A=
9; + B&#39; + C&#39;.<br><br>&gt; If D&#39; is higher-fee than P1 or P2 or =
P3 but inferior to the sum of HTLCs confirmed by P1+P2+P3, I think it&#39;s=
 lucrative for the attacker ?<br><br>I could be misunderstanding, but an at=
tacker wouldn&#39;t be able to batch-attack like this. Alice&#39;s package =
only conflicts with A&#39; + D&#39;, not A&#39; + B&#39; + C&#39; + D&#39;.=
 She only needs to pay for evicting 2 transactions.<br><br>&gt; Do we assum=
e that broadcasted packages are &quot;honest&quot; by default and that the =
parent(s) always need the child to pass the fee checks, that way saving the=
 processing of individual transactions which are expected to fail in 99% of=
 cases or more ad hoc composition of packages at relay ?<br>&gt; I think th=
is point is quite dependent on the p2p packages format/logic we&#39;ll end =
up on and that we should feel free to revisit it later ?<br><br>I think it&=
#39;s the opposite; there&#39;s no way for us to assume that p2p packages w=
ill be &quot;honest.&quot; I&#39;d like to have two things before we expose=
 on P2P: (1) ensure that the amount of resources potentially allocated for =
package validation isn&#39;t disproportionately higher than that of single =
transaction validation and (2) only use package validation when we&#39;re u=
nsatisifed with the single validation result, e.g. we might get better fees=
.<br>Yes, let&#39;s revisit this later :)<br>=C2=A0<br>=C2=A0&gt; Yes, if y=
ou receive A+B, and A is already in-mempoo, I agree you can discard its fee=
rate as B should pay for all fees checked on its own. Where I&#39;m unclear=
 is when you have in-mempool A+B and receive A+B&#39;. Should B&#39; have a=
 fee high enough to cover the bandwidth penalty replacement (`PaysForRBF`, =
2nd check) of both A+B&#39; or only B&#39; ?<br>=C2=A0<br>=C2=A0B&#39; only=
 needs to pay for itself in this case.<br>=C2=A0<br>&gt; &gt; Do we want th=
e child to be able to replace mempool transactions as well?<br><br>&gt; If =
we mean when you have replaceable A+B then A&#39;+B&#39; try to replace wit=
h a higher-feerate ? I think that&#39;s exactly the case we need for Lightn=
ing as A+B is coming from Alice and A&#39;+B&#39; is coming from Bob :/<br>=
<br><div>Let me clarify this because I can see that my wording was ambiguou=
s, and then please let me know if it fits Lightning&#39;s needs?</div><div>=
<br></div>In my proposal, I wrote &quot;If a package meets feerate requirem=
ents as a package, the parents in the transaction are allowed to replace-by=
-fee mempool transactions. The child cannot replace mempool transactions.&q=
uot; What I meant was: the package can replace mempool transactions if any =
of the parents conflict with mempool transactions. The child cannot not con=
flict with any mempool transactions.<br>The Lightning use case this attempt=
s to address is: Alice and Mallory are LN counterparties, and have packages=
 A+B and A&#39;+B&#39;, respectively. A and A&#39; are their commitment tra=
nsactions and conflict with each other; they have shared inputs and differe=
nt txids.<br>B spends Alice&#39;s anchor output from A. B&#39; spends Mallo=
ry&#39;s anchor output from A&#39;. Thus, B and B&#39; do not conflict with=
 each other.<br>Alice can broadcast her package, A+B, to replace Mallory=
9;s package, A&#39;+B&#39;, since B doesn&#39;t conflict with the mempool.<=
br><br>Would this be ok?<br><br>&gt; The second option, a child of A&#39;, =
In the LN case I think the CPFP is attached on one&#39;s anchor output.<br>=
<br>While it would be nice to have full RBF, malleability of the child won&=
#39;t block RBF here. If we&#39;re trying to replace A&#39;, we only requir=
e that A&#39; signals replaceability, and don&#39;t mind if its child doesn=
&#39;t.<br><br>&gt; &gt; B has an ancestor score of 10sat/vb and D has an<b=
r>&gt; &gt; ancestor score of ~2.9sat/vb. Since D&#39;s ancestor score is l=
ower than B&#39;s,<br>&gt; &gt; it fails the proposed package RBF Rule #2, =
so this package would be<br>&gt; &gt; rejected. Does this meet your expecta=
tions?<br><br>&gt; Well what sounds odd to me, in my example, we fail D eve=
n if it has a higher-fee than B. Like A+B absolute fees are 2000 sats and A=
+C+D absolute fees are 4500 sats ?<br><br>Yes, A+C+D pays 2500sat more in f=
ees, but it is also 1000vB larger. A miner should prefer to utilize their b=
lock space more effectively.<br><br>&gt; Is this compatible with a model wh=
ere a miner prioritizes absolute fees over ancestor score, in the case that=
 mempools aren&#39;t full-enough to fulfill a block ?<br><br>No, because we=
 don&#39;t use that model.<br><br>Thanks,<br>Gloria</div><br><div class=3D"=
gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Thu, Sep 23, 2021 at =
5:29 AM Antoine Riard &lt;<a href=3D"mailto:antoine.riard@gmail.com" target=
=3D"_blank">antoine.riard@gmail.com</a>&gt; wrote:<br></div><blockquote cla=
ss=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid =
rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr">&gt; Correct, if B+C is=
 too low feerate to be accepted, we will reject it. I<br>&gt; prefer this b=
ecause it is incentive compatible: A can be mined by itself,<br>&gt; so the=
re&#39;s no reason to prefer A+B+C instead of A.<br>&gt; As another way of =
looking at this, consider the case where we do accept<br>&gt; A+B+C and it =
sits at the &quot;bottom&quot; of our mempool. If our mempool reaches<br>&g=
t; capacity, we evict the lowest descendant feerate transactions, which are=
<br>&gt; B+C in this case. This gives us the same resulting mempool, with A=
 and not<br>&gt; B+C.<br><br>I agree here. Doing otherwise, we might evict =
other transactions mempool in `MempoolAccept::Finalize` with a higher-feera=
te than B+C while those evicted transactions are the most compelling for bl=
ock construction.<br><br>I thought at first missing this acceptance require=
ment would break a fee-bumping scheme like Parent-Pay-For-Child where a hig=
h-fee parent is attached to a child signed with SIGHASH_ANYONECANPAY but in=
 this case the child fee is capturing the parent value. I can&#39;t think o=
f other fee-bumping schemes potentially affected. If they do exist I would =
say they&#39;re wrong in their design assumptions.<br><br>&gt; If or when w=
e have witness replacement, the logic is: if the individual<br>&gt; transac=
tion is enough to replace the mempool one, the replacement will<br>&gt; hap=
pen during the preceding individual transaction acceptance, and<br>&gt; ded=
uplication logic will work. Otherwise, we will try to deduplicate by<br>&gt=
; wtxid, see that we need a package witness replacement, and use the packag=
e<br>&gt; feerate to evaluate whether this is economically rational.<br><br=
>IIUC, you have package A+B, during the dedup phase early in `AcceptMultipl=
eTransactions` if you observe same-txid-different-wtixd A&#39; and A&#39; i=
s higher feerate than A, you trim A and replace by A&#39; ?<br><br>I think =
this approach is safe, the one who appears unsafe to me is when A&#39; has =
a _lower_ feerate, even if A&#39; is already accepted by our mempool ? In t=
hat case iirc that would be a pinning.<br><br>Good to see progress on witne=
ss replacement before we see usage of Taproot tree in the context of multi-=
party, where a malicious counterparty inflates its witness to jam a honest =
spending.<br><br>(Note, the commit linked currently points nowhere :))<br><=
br><br>&gt; Please note that A may replace A&#39; even if A&#39; has higher=
 fees than A<br>&gt; individually, because the proposed package RBF utilize=
s the fees and size<br>&gt; of the entire package. This just requires E to =
pay enough fees, although<br>&gt; this can be pretty high if there are also=
 potential B&#39; and C&#39; competing<br>&gt; commitment transactions that=
 we don&#39;t know about.<br><br>Ah right, if the package acceptance waives=
 `PaysMoreThanConflicts` for the individual check on A, the honest package =
should replace the pinning attempt. I&#39;ve not fully parsed the proposed =
implementation yet.<br><br>Though note, I think it&#39;s still unsafe for a=
 Lightning multi-commitment-broadcast-as-one-package as a malicious A&#39; =
might have an absolute fee higher than E. It sounds uneconomical for<br>an =
attacker but I think it&#39;s not when you consider than you can &quot;batc=
h&quot; attack against multiple honest counterparties. E.g, Mallory broadca=
st A&#39; + B&#39; + C&#39; + D&#39; where A&#39; conflicts with Alice&#39;=
s honest package P1, B&#39; conflicts with Bob&#39;s honest package P2, C&#=
39; conflicts with Caroll&#39;s honest package P3. And D&#39; is a high-fee=
 child of A&#39; + B&#39; + C&#39;.<br><br>If D&#39; is higher-fee than P1 =
or P2 or P3 but inferior to the sum of HTLCs confirmed by P1+P2+P3, I think=
 it&#39;s lucrative for the attacker ?<br><br>&gt; So far, my understanding=
 is that multi-parent-1-child is desired for<br>&gt; batched fee-bumping (<=
br>&gt; <a href=3D"https://github.com/bitcoin/bitcoin/pull/22674#issuecomme=
nt-897951289" target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/226=
74#issuecomment-897951289</a>) and<br>&gt; I&#39;ve also seen your response=
 which I have less context on (<br>&gt; <a href=3D"https://github.com/bitco=
in/bitcoin/pull/22674#issuecomment-900352202" target=3D"_blank">https://git=
hub.com/bitcoin/bitcoin/pull/22674#issuecomment-900352202</a>). That<br>&gt=
; being said, I am happy to create a new proposal for 1 parent + 1 child<br=
>&gt; (which would be slightly simpler) and plan for moving to<br>&gt; mult=
i-parent-1-child later if that is preferred. I am very interested in<br>&gt=
; hearing feedback on that approach.<br><br>I think batched fee-bumping is =
okay as long as you don&#39;t have time-sensitive outputs encumbering your =
commitment transactions. For the reasons mentioned above, I think that&#39;=
s unsafe.<br><br>What I&#39;m worried about is=C2=A0 L2 developers, potenti=
ally not aware about all the mempool subtleties blurring the difference and=
 always batching their broadcast by default.<br><br>IMO, a good thing by re=
straining to 1-parent + 1 child, =C2=A0we artificially constraint L2 design=
 space for now and minimize risks of unsafe usage of the package API :)<br>=
<br>I think that&#39;s a point where it would be relevant to have the opini=
on of more L2 devs.<br><br>&gt; I think there is a misunderstanding here - =
let me describe what I&#39;m<br>&gt; proposing we&#39;d do in this situatio=
n: we&#39;ll try individual submission for A,<br>&gt; see that it fails due=
 to &quot;insufficient fees.&quot; Then, we&#39;ll try package<br>&gt; vali=
dation for A+B and use package RBF. If A+B pays enough, it can still<br>&gt=
; replace A&#39;. If A fails for a bad signature, we won&#39;t look at B or=
 A+B. Does<br>&gt; this meet your expectations?<br><br>Yes there was a misu=
nderstanding, I think this approach is correct, it&#39;s more a question of=
 performance. Do we assume that broadcasted packages are &quot;honest&quot;=
 by default and that the parent(s) always need the child to pass the fee ch=
ecks, that way saving the processing of individual transactions which are e=
xpected to fail in 99% of cases or more ad hoc composition of packages at r=
elay ?<br><br>I think this point is quite dependent on the p2p packages for=
mat/logic we&#39;ll end up on and that we should feel free to revisit it la=
ter ?<br><br><br>&gt; What problem are you trying to solve by the package f=
eerate *after* dedup<br>rule ?<br>&gt; My understanding is that an in-packa=
ge transaction might be already in<br>the mempool. Therefore, to compute a =
correct RBF penalty replacement, the<br>vsize of this transaction could be =
discarded lowering the cost of package<br>RBF.<br><br>&gt; I&#39;m proposin=
g that, when a transaction has already been submitted to<br>&gt; mempool, w=
e would ignore both its fees and vsize when calculating package<br>&gt; fee=
rate. <br><br>Yes, if you receive A+B, and A is already in-mempoo, I agree =
you can discard its feerate as B should pay for all fees checked on its own=
. Where I&#39;m unclear is when you have in-mempool A+B and receive A+B&#39=
;. Should B&#39; have a fee high enough to cover the bandwidth penalty repl=
acement (`PaysForRBF`, 2nd check) of both A+B&#39; or only B&#39; ?<br><br>=
If you have a second-layer like current Lightning, you might have a counter=
party commitment to replace and should always expect to have to pay for par=
ent replacement bandwidth.<br><br>Where a potential discount sounds interes=
ting is when you have an univoque state on the first-stage of transactions.=
 E.g DLC&#39;s funding transaction which might be CPFP by any participant i=
irc.<br><br>&gt; Note that, if C&#39; conflicts with C, it also conflicts w=
ith D, since D is a<br>&gt; descendant of C and would thus need to be evict=
ed along with it.<br><br>Ah once again I think it&#39;s a misunderstanding =
without the code under my eyes! If we do C&#39; `PreChecks`, solve the conf=
licts provoked by it, i.e mark for potential eviction D and don&#39;t consi=
der it for future conflicts in the rest of the package, I think D&#39; `Pre=
Checks` should be good ?<br><br>&gt; More generally, this example is surpri=
sing to me because I didn&#39;t think<br>&gt; packages would be used to fee=
-bump replaceable transactions. Do we want the<br>&gt; child to be able to =
replace mempool transactions as well?<br><br>If we mean when you have repla=
ceable A+B then A&#39;+B&#39; try to replace with a higher-feerate ? I thin=
k that&#39;s exactly the case we need for Lightning as A+B is coming from A=
lice and A&#39;+B&#39; is coming from Bob :/<br><br>&gt; I&#39;m not sure w=
hat you mean? Let&#39;s say we have a package of parent A + child<br>&gt; B=
, where A is supposed to replace a mempool transaction A&#39;. Are you sayi=
ng<br>&gt; that counterparties are able to malleate the package child B, or=
 a child of<br>&gt; A&#39;? <br><br>The second option, a child of A&#39;, I=
n the LN case I think the CPFP is attached on one&#39;s anchor output.<br><=
br>I think it&#39;s good if we assume the solve-conflicts-after-parent&#39;=
s`&#39;PreChecks` mentioned above or fixing inherited signaling or full-rbf=
 ?<br><br>&gt; Sorry, I don&#39;t understand what you mean by &quot;preserv=
e the package<br>&gt; integrity?&quot; Could you elaborate?<br><br>After th=
inking the relaxation about the &quot;new&quot; unconfirmed input is not li=
nked to trimming but I would say more to the multi-parent support.<br><br>L=
et&#39;s say you have A+B trying to replace C+D where B is also spending al=
ready in-mempool E. To succeed, you need to waive the no-new-unconfirmed in=
put as D isn&#39;t spending E.<br><br>So good, I think we agree on the prob=
lem description here.<br><br>&gt; I am in agreement with your calculations =
but unsure if we disagree on the<br>&gt; expected outcome. Yes, B has an an=
cestor score of 10sat/vb and D has an<br>&gt; ancestor score of ~2.9sat/vb.=
 Since D&#39;s ancestor score is lower than B&#39;s,<br>&gt; it fails the p=
roposed package RBF Rule #2, so this package would be<br>&gt; rejected. Doe=
s this meet your expectations?<br><br>Well what sounds odd to me, in my exa=
mple, we fail D even if it has a higher-fee than B. Like A+B absolute fees =
are 2000 sats and A+C+D absolute fees are 4500 sats ?<br><br>Is this compat=
ible with a model where a miner prioritizes absolute fees over ancestor sco=
re, in the case that mempools aren&#39;t full-enough to fulfill a block ?<b=
r><br>Let me know if I can clarify a point.<br><br>Antoine<br></div><br><di=
v class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">Le=C2=A0lun. =
20 sept. 2021 =C3=A0=C2=A011:10, Gloria Zhao &lt;<a href=3D"mailto:gloriajz=
hao@gmail.com" target=3D"_blank">gloriajzhao@gmail.com</a>&gt; a =C3=A9crit=
=C2=A0:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px =
0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=
=3D"ltr"><br>Hi Antoine,<br><br>First of all, thank you for the thorough re=
view. I appreciate your insight on LN requirements.<br><br>&gt; IIUC, you h=
ave a package A+B+C submitted for acceptance and A is already in your mempo=
ol. You trim out A from the package and then evaluate B+C.<br><br>&gt; I th=
ink this might be an issue if A is the higher-fee element of the ABC packag=
e. B+C package fees might be under the mempool min fee and will be rejected=
, potentially breaking the acceptance expectations of the package issuer ?<=
br><br>Correct, if B+C is too low feerate to be accepted, we will reject it=
. I prefer this because it is incentive compatible: A can be mined by itsel=
f, so there&#39;s no reason to prefer A+B+C instead of A.<br>As another way=
 of looking at this, consider the case where we do accept A+B+C and it sits=
 at the &quot;bottom&quot; of our mempool. If our mempool reaches capacity,=
 we evict the lowest descendant feerate transactions, which are B+C in this=
 case. This gives us the same resulting mempool, with A and not B+C.<br><di=
v><br></div><div><br></div>&gt; Further, I think the dedup should be done o=
n wtxid, as you might have multiple valid witnesses. Though with varying vs=
izes and as such offering different feerates.<br><br>I agree that variation=
s of the same package with different witnesses is a case that must be handl=
ed. I consider witness replacement to be a project that can be done in para=
llel to package mempool acceptance because being able to accept packages do=
es not worsen the problem of a same-txid-different-witness &quot;pinning&qu=
ot; attack.<br><br>If or when we have witness replacement, the logic is: if=
 the individual transaction is enough to replace the mempool one, the repla=
cement will happen during the preceding individual transaction acceptance, =
and deduplication logic will work. Otherwise, we will try to deduplicate by=
 wtxid, see that we need a package witness replacement, and use the package=
 feerate to evaluate whether this is economically rational.<br><br>See the =
#22290 &quot;handle package transactions already in mempool&quot; commit (<=
a href=3D"https://github.com/bitcoin/bitcoin/pull/22290/commits/fea75a2237b=
46cf76145242fecad7e274bfcb5ff" target=3D"_blank">https://github.com/bitcoin=
/bitcoin/pull/22290/commits/fea75a2237b46cf76145242fecad7e274bfcb5ff</a>), =
which handles the case of same-txid-different-witness by simply using the t=
ransaction in the mempool for now, with TODOs for what I just described.<br=
><br><br>&gt; I&#39;m not clearly understanding the accepted topologies. By=
 &quot;parent and child to share a parent&quot;, do you mean the set of tra=
nsactions A, B, C, where B is spending A and C is spending A and B would be=
 correct ?<br><br>Yes, that is what I meant. Yes, that would a valid packag=
e under these rules.<br><br>&gt; If yes, is there a width-limit introduced =
or we fallback on MAX_PACKAGE_COUNT=3D25 ?<br><br>No, there is no limit on =
connectivity other than &quot;child with all unconfirmed parents.&quot; We =
will enforce MAX_PACKAGE_COUNT=3D25 and child&#39;s in-mempool + in-package=
 ancestor limits.<br><br><br>&gt; Considering the current Core&#39;s mempoo=
l acceptance rules, I think CPFP batching is unsafe for LN time-sensitive c=
losure. A malicious tx-relay jamming successful on one channel commitment t=
ransaction would contamine the remaining commitments sharing the same packa=
ge.<br><br>&gt; E.g, you broadcast the package A+B+C+D+E where A,B,C,D are =
commitment transactions and E a shared CPFP. If a malicious A&#39; transact=
ion has a better feerate than A, the whole package acceptance will fail. Ev=
en if A&#39; confirms in the following block,<br>the propagation and confir=
mation of B+C+D have been delayed. This could carry on a loss of funds.<br>=
<br>Please note that A may replace A&#39; even if A&#39; has higher fees th=
an A individually, because the proposed package RBF utilizes the fees and s=
ize of the entire package. This just requires E to pay enough fees, althoug=
h this can be pretty high if there are also potential B&#39; and C&#39; com=
peting commitment transactions that we don&#39;t know about.<br><br><br>&gt=
; IMHO, I&#39;m leaning towards deploying during a first phase 1-parent/1-c=
hild. I think it&#39;s the most conservative step still improving second-la=
yer safety.<br><br>So far, my understanding is that multi-parent-1-child is=
 desired for batched fee-bumping (<a href=3D"https://github.com/bitcoin/bit=
coin/pull/22674#issuecomment-897951289" target=3D"_blank">https://github.co=
m/bitcoin/bitcoin/pull/22674#issuecomment-897951289</a>) and I&#39;ve also =
seen your response which I have less context on (<a href=3D"https://github.=
com/bitcoin/bitcoin/pull/22674#issuecomment-900352202" target=3D"_blank">ht=
tps://github.com/bitcoin/bitcoin/pull/22674#issuecomment-900352202</a>). Th=
at being said, I am happy to create a new proposal for 1 parent + 1 child (=
which would be slightly simpler) and plan for moving to multi-parent-1-chil=
d later if that is preferred. I am very interested in hearing feedback on t=
hat approach.<br><br><br>&gt; If A+B is submitted to replace A&#39;, where =
A pays 0 sats, B pays 200 sats and A&#39; pays 100 sats. If we apply the in=
dividual RBF on A, A+B acceptance fails. For this reason I think the indivi=
dual RBF should be bypassed and only the package RBF apply ?<br><br>I think=
 there is a misunderstanding here - let me describe what I&#39;m proposing =
we&#39;d do in this situation: we&#39;ll try individual submission for A, s=
ee that it fails due to &quot;insufficient fees.&quot; Then, we&#39;ll try =
package validation for A+B and use package RBF. If A+B pays enough, it can =
still replace A&#39;. If A fails for a bad signature, we won&#39;t look at =
B or A+B. Does this meet your expectations?<br><br><br>&gt; What problem ar=
e you trying to solve by the package feerate *after* dedup rule ?<br>&gt; M=
y understanding is that an in-package transaction might be already in the m=
empool. Therefore, to compute a correct RBF penalty replacement, the vsize =
of this transaction could be discarded lowering the cost of package RBF.<br=
><br>I&#39;m proposing that, when a transaction has already been submitted =
to mempool, we would ignore both its fees and vsize when calculating packag=
e feerate. In example G2, we shouldn&#39;t count M1 fees after its submissi=
on to mempool, since M1&#39;s fees have already been used to pay for its in=
dividual bandwidth, and it shouldn&#39;t be used again to pay for P2 and P3=
&#39;s bandwidth. We also shouldn&#39;t count its vsize, since it has alrea=
dy been paid for.<br><br><br>&gt; I think this is a footgunish API, as if a=
 package issuer send the multiple-parent-one-child package A,B,C,D where D =
is the child of A,B,C. Then try to broadcast the higher-feerate C&#39;+D=
9; package, it should be rejected. So it&#39;s breaking the naive broadcast=
er assumption that a higher-feerate/higher-fee package always replaces ?<br=
><br><div>Note that, if C&#39; conflicts with C, it also conflicts with D, =
since D is a descendant of C and would thus need to be evicted along with i=
t. Implicitly, D&#39; would not be in conflict with D.<br></div><div>More g=
enerally, this example is surprising to me because I didn&#39;t think packa=
ges would be used to fee-bump replaceable transactions. Do we want the chil=
d to be able to replace mempool transactions as well? This can be implement=
ed with a bit of additional logic.</div><br>&gt; I think this is unsafe for=
 L2s if counterparties have malleability of the child transaction. They can=
 block your package replacement by opting-out from RBF signaling. IIRC, LN&=
#39;s &quot;anchor output&quot; presents such an ability.<br><br>I&#39;m no=
t sure what you mean? Let&#39;s say we have a package of parent A + child B=
, where A is supposed to replace a mempool transaction A&#39;. Are you sayi=
ng that counterparties are able to malleate the package child B, or a child=
 of A&#39;? If they can malleate a child of A&#39;, that shouldn&#39;t matt=
er as long as A&#39; is signaling replacement. This would be handled identi=
cally with full RBF and what Core currently implements.<br><br>&gt; I think=
 this is an issue brought by the trimming during the dedup phase. If we pre=
serve the package integrity, only re-using the tx-level checks results of a=
lready in-mempool transactions to gain in CPU time we won&#39;t have this i=
ssue. Package childs can add unconfirmed inputs as long as they&#39;re in-p=
ackage, the bip125 rule2 is only evaluated against parents ?<br><br>Sorry, =
I don&#39;t understand what you mean by &quot;preserve the package integrit=
y?&quot; Could you elaborate?<br><br>&gt; Let&#39;s say you have in-mempool=
 A, B where A pays 10 sat/vb for 100 vbytes and B pays 10 sat/vb for 100 vb=
ytes. You have the candidate replacement D spending both A and C where D pa=
ys 15sat/vb for 100 vbytes and C pays 1 sat/vb for 1000 vbytes.<br><br>&gt;=
 Package A + B ancestor score is 10 sat/vb.<br><br>&gt; D has a higher feer=
ate/absolute fee than B.<br><br>&gt; Package A + C + D ancestor score is ~ =
3 sat/vb ((A&#39;s 1000 sats + C&#39;s 1000 sats + D&#39;s 1500 sats) / A&#=
39;s 100 vb + C&#39;s 1000 vb + D&#39;s 100 vb)<br><br><div>I am in agreeme=
nt with your calculations but unsure if we disagree on the expected outcome=
. Yes, B has an ancestor score of 10sat/vb and D has an ancestor score of ~=
2.9sat/vb. Since D&#39;s ancestor score is lower than B&#39;s, it fails the=
 proposed package RBF Rule #2, so this package would be rejected. Does this=
 meet your expectations?</div><div><br></div><div>Thank you for linking to =
projects that might be interested in package relay :)<br></div><br>Thanks,<=
br>Gloria<br></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D=
"gmail_attr">On Mon, Sep 20, 2021 at 12:16 AM Antoine Riard &lt;<a href=3D"=
mailto:antoine.riard@gmail.com" target=3D"_blank">antoine.riard@gmail.com</=
a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0p=
x 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><d=
iv dir=3D"ltr">Hi Gloria,<br><br>&gt; A package may contain transactions th=
at are already in the mempool. We<br>&gt; remove<br>&gt; (&quot;deduplicate=
&quot;) those transactions from the package for the purposes of<br>&gt; pac=
kage<br>&gt; mempool acceptance. If a package is empty after deduplication,=
 we do<br>&gt; nothing.<br><br>IIUC, you have a package A+B+C submitted for=
 acceptance and A is already in your mempool. You trim out A from the packa=
ge and then evaluate B+C.<br><br>I think this might be an issue if A is the=
 higher-fee element of the ABC package. B+C package fees might be under the=
 mempool min fee and will be rejected, potentially breaking the acceptance =
expectations of the package issuer ?<br><br>Further, I think the dedup shou=
ld be done on wtxid, as you might have multiple valid witnesses. Though wit=
h varying vsizes and as such offering different feerates.<br><br>E.g you=
9;re going to evaluate the package A+B and A&#39; is already in your mempoo=
l with a bigger valid witness. You trim A based on txid, then you evaluate =
A&#39;+B, which fails the fee checks. However, evaluating A+B would have be=
en a success.<br><br>AFAICT, the dedup rationale would be to save on CPU ti=
me/IO disk, to avoid repeated signatures verification and parent UTXOs fetc=
hes ? Can we achieve the same goal by bypassing tx-level checks for already=
-in txn while conserving the package integrity for package-level checks ?<b=
r><br>&gt; Note that it&#39;s possible for the parents to be<br>&gt; indire=
ct<br>&gt; descendants/ancestors of one another, or for parent and child to=
 share a<br>&gt; parent,<br>&gt; so we cannot make any other topology assum=
ptions.<br><br>I&#39;m not clearly understanding the accepted topologies. B=
y &quot;parent and child to share a parent&quot;, do you mean the set of tr=
ansactions A, B, C, where B is spending A and C is spending A and B would b=
e correct ?<br><br>If yes, is there a width-limit introduced or we fallback=
 on MAX_PACKAGE_COUNT=3D25 ?<br><br>IIRC, one rationale to come with this t=
opology limitation was to lower the DoS risks when potentially deploying p2=
p packages.<br><br>Considering the current Core&#39;s mempool acceptance ru=
les, I think CPFP batching is unsafe for LN time-sensitive closure. A malic=
ious tx-relay jamming successful on one channel commitment transaction woul=
d contamine the remaining commitments sharing the same package.<br><br>E.g,=
 you broadcast the package A+B+C+D+E where A,B,C,D are commitment transacti=
ons and E a shared CPFP. If a malicious A&#39; transaction has a better fee=
rate than A, the whole package acceptance will fail. Even if A&#39; confirm=
s in the following block, <br>the propagation and confirmation of B+C+D hav=
e been delayed. This could carry on a loss of funds.<br><br>That said, if y=
ou&#39;re broadcasting commitment transactions without time-sensitive HTLC =
outputs, I think the batching is effectively a fee saving as you don&#39;t =
have to duplicate the CPFP.<br><br>IMHO, I&#39;m leaning towards deploying =
during a first phase 1-parent/1-child. I think it&#39;s the most conservati=
ve step still improving second-layer safety.<br><br>&gt; *Rationale*: =C2=
=A0It would be incorrect to use the fees of transactions that are<br>&gt; a=
lready in the mempool, as we do not want a transaction&#39;s fees to be<br>=
&gt; double-counted for both its individual RBF and package RBF.<br><br>I&#=
39;m unsure about the logical order of the checks proposed.<br><br>If A+B i=
s submitted to replace A&#39;, where A pays 0 sats, B pays 200 sats and A&#=
39; pays 100 sats. If we apply the individual RBF on A, A+B acceptance fail=
s. For this reason I think the individual RBF should be bypassed and only t=
he package RBF apply ? <br><br>Note this situation is plausible, with curre=
nt LN design, your counterparty can have a commitment transaction with a be=
tter fee just by selecting a higher `dust_limit_satoshis` than yours.<br><b=
r>&gt; Examples F and G [14] show the same package, but P1 is submitted<br>=
&gt; individually before<br>&gt; the package in example G. In example F, we=
 can see that the 300vB package<br>&gt; pays<br>&gt; an additional 200sat i=
n fees, which is not enough to pay for its own<br>&gt; bandwidth<br>&gt; (B=
IP125#4). In example G, we can see that P1 pays enough to replace M1, but<b=
r>&gt; using P1&#39;s fees again during package submission would make it lo=
ok like a<br>&gt; 300sat<br>&gt; increase for a 200vB package. Even includi=
ng its fees and size would not be<br>&gt; sufficient in this example, since=
 the 300sat looks like enough for the 300vB<br>&gt; package. The calculcati=
on after deduplication is 100sat increase for a<br>&gt; package<br>&gt; of =
size 200vB, which correctly fails BIP125#4. Assume all transactions have<br=
>&gt; a<br>&gt; size of 100vB.<br><br>What problem are you trying to solve =
by the package feerate *after* dedup rule ?<br><br>My understanding is that=
 an in-package transaction might be already in the mempool. Therefore, to c=
ompute a correct RBF penalty replacement, the vsize of this transaction cou=
ld be discarded lowering the cost of package RBF.<br><br>If we keep a &quot=
;safe&quot; dedup mechanism (see my point above), I think this discount is =
justified, as the validation cost of node operators is paid for ?<br><br>&g=
t; The child cannot replace mempool transactions.<br><br>Let&#39;s say you =
issue package A+B, then package C+B&#39;, where B&#39; is a child of both A=
 and C. This rule fails the acceptance of C+B&#39; ?<br><br>I think this is=
 a footgunish API, as if a package issuer send the multiple-parent-one-chil=
d package A,B,C,D where D is the child of A,B,C. Then try to broadcast the =
higher-feerate C&#39;+D&#39; package, it should be rejected. So it&#39;s br=
eaking the naive broadcaster assumption that a higher-feerate/higher-fee pa=
ckage always replaces ? And it might be unsafe in protocols where states ar=
e symmetric. E.g a malicious counterparty broadcasts first S+A, then you ho=
nestly broadcast S+B, where B pays better fees.<br><br>&gt; All mempool tra=
nsactions to be replaced must signal replaceability.<br><br>I think this is=
 unsafe for L2s if counterparties have malleability of the child transactio=
n. They can block your package replacement by opting-out from RBF signaling=
. IIRC, LN&#39;s &quot;anchor output&quot; presents such an ability.<br><br=
>I think it&#39;s better to either fix inherited signaling or move towards =
full-rbf.<br><br>&gt; if a package parent has already been submitted, it wo=
uld<br>&gt; look<br>&gt;like the child is spending a &quot;new&quot; unconf=
irmed input.<br><br>I think this is an issue brought by the trimming during=
 the dedup phase. If we preserve the package integrity, only re-using the t=
x-level checks results of already in-mempool transactions to gain in CPU ti=
me we won&#39;t have this issue. Package childs can add unconfirmed inputs =
as long as they&#39;re in-package, the bip125 rule2 is only evaluated again=
st parents ?<br><br>&gt; However, we still achieve the same goal of requiri=
ng the<br>&gt; replacement<br>&gt; transactions to have a ancestor score at=
 least as high as the original<br>&gt; ones.<br><br>I&#39;m not sure if thi=
s holds...<br><br>Let&#39;s say you have in-mempool A, B where A pays 10 sa=
t/vb for 100 vbytes and B pays 10 sat/vb for 100 vbytes. You have the candi=
date replacement D spending both A and C where D pays 15sat/vb for 100 vbyt=
es and C pays 1 sat/vb for 1000 vbytes.<br><br>Package A + B ancestor score=
 is 10 sat/vb.<br><br>D has a higher feerate/absolute fee than B.<br><br>Pa=
ckage A + C + D ancestor score is ~ 3 sat/vb ((A&#39;s 1000 sats + C&#39;s =
1000 sats + D&#39;s 1500 sats) / <br>A&#39;s 100 vb + C&#39;s 1000 vb + D&#=
39;s 100 vb)<br><br>Overall, this is a review through the lenses of LN requ=
irements. I think other L2 protocols/applications<br>could be candidates to=
 using package accept/relay such as:<br>* <a href=3D"https://github.com/lig=
htninglabs/pool" target=3D"_blank">https://github.com/lightninglabs/pool</a=
><br>* <a href=3D"https://github.com/discreetlogcontracts/dlcspecs" target=
=3D"_blank">https://github.com/discreetlogcontracts/dlcspecs</a><br>* <a hr=
ef=3D"https://github.com/bitcoin-teleport/teleport-transactions/" target=3D=
"_blank">https://github.com/bitcoin-teleport/teleport-transactions/</a><br>=
* <a href=3D"https://github.com/sapio-lang/sapio" target=3D"_blank">https:/=
/github.com/sapio-lang/sapio</a><br>* <a href=3D"https://github.com/commerc=
eblock/mercury/blob/master/doc/statechains.md" target=3D"_blank">https://gi=
thub.com/commerceblock/mercury/blob/master/doc/statechains.md</a><br>* <a h=
ref=3D"https://github.com/revault/practical-revault" target=3D"_blank">http=
s://github.com/revault/practical-revault</a><br><br>Thanks for rolling forw=
ard the ball on this subject.<br><br>Antoine<br></div><br><div class=3D"gma=
il_quote"><div dir=3D"ltr" class=3D"gmail_attr">Le=C2=A0jeu. 16 sept. 2021 =
=C3=A0=C2=A003:55, Gloria Zhao via bitcoin-dev &lt;<a href=3D"mailto:bitcoi=
n-dev@lists.linuxfoundation.org" target=3D"_blank">bitcoin-dev@lists.linuxf=
oundation.org</a>&gt; a =C3=A9crit=C2=A0:<br></div><blockquote class=3D"gma=
il_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,2=
04,204);padding-left:1ex"><div dir=3D"ltr">Hi there,<br><br>I&#39;m writing=
 to propose a set of mempool policy changes to enable package<br>validation=
 (in preparation for package relay) in Bitcoin Core. These would not<br>be =
consensus or P2P protocol changes. However, since mempool policy<br>signifi=
cantly affects transaction propagation, I believe this is relevant for<br>t=
he mailing list.<br><br>My proposal enables packages consisting of multiple=
 parents and 1 child. If you<br>develop software that relies on specific tr=
ansaction relay assumptions and/or<br>are interested in using package relay=
 in the future, I&#39;m very interested to hear<br>your feedback on the uti=
lity or restrictiveness of these package policies for<br>your use cases.<br=
><br>A draft implementation of this proposal can be found in [Bitcoin Core<=
br>PR#22290][1].<br><br>An illustrated version of this post can be found at=
<br><div><a href=3D"https://gist.github.com/glozow/dc4e9d5c5b14ade7cdfac40f=
43adb18a" target=3D"_blank">https://gist.github.com/glozow/dc4e9d5c5b14ade7=
cdfac40f43adb18a</a>.</div><div>I have also linked the images below.</div><=
br>## Background<br><br>Feel free to skip this section if you are already f=
amiliar with mempool policy<br>and package relay terminology.<br><br>### Te=
rminology Clarifications<br><br>* Package =3D an ordered list of related tr=
ansactions, representable by a Directed<br>=C2=A0 Acyclic Graph.<br>* Packa=
ge Feerate =3D the total modified fees divided by the total virtual size of=
<br>=C2=A0 all transactions in the package.<br>=C2=A0 =C2=A0 - Modified fee=
s =3D a transaction&#39;s base fees + fee delta applied by the user<br>=C2=
=A0 =C2=A0 =C2=A0 with `prioritisetransaction`. As such, we expect this to =
vary across<br>mempools.<br>=C2=A0 =C2=A0 - Virtual Size =3D the maximum of=
 virtual sizes calculated using [BIP141<br>=C2=A0 =C2=A0 =C2=A0 virtual siz=
e][2] and sigop weight. [Implemented here in Bitcoin Core][3].<br>=C2=A0 =
=C2=A0 - Note that feerate is not necessarily based on the base fees and se=
rialized<br>=C2=A0 =C2=A0 =C2=A0 size.<br><br>* Fee-Bumping =3D user/wallet=
 actions that take advantage of miner incentives to<br>=C2=A0 boost a trans=
action&#39;s candidacy for inclusion in a block, including Child Pays<br>fo=
r Parent (CPFP) and [BIP125][12] Replace-by-Fee (RBF). Our intention in<br>=
mempool policy is to recognize when the new transaction is more economical =
to<br>mine than the original one(s) but not open DoS vectors, so there are =
some<br>limitations.<br><br>### Policy<br><br>The purpose of the mempool is=
 to store the best (to be most incentive-compatible<br>with miners, highest=
 feerate) candidates for inclusion in a block. Miners use<br>the mempool to=
 build block templates. The mempool is also useful as a cache for<br>boosti=
ng block relay and validation performance, aiding transaction relay, and<br=
>generating feerate estimations.<br><br>Ideally, all consensus-valid transa=
ctions paying reasonable fees should make it<br>to miners through normal tr=
ansaction relay, without any special connectivity or<br>relationships with =
miners. On the other hand, nodes do not have unlimited<br>resources, and a =
P2P network designed to let any honest node broadcast their<br>transactions=
 also exposes the transaction validation engine to DoS attacks from<br>mali=
cious peers.<br><br>As such, for unconfirmed transactions we are considerin=
g for our mempool, we<br>apply a set of validation rules in addition to con=
sensus, primarily to protect<br>us from resource exhaustion and aid our eff=
orts to keep the highest fee<br>transactions. We call this mempool _policy_=
: a set of (configurable,<br>node-specific) rules that transactions must ab=
ide by in order to be accepted<br>into our mempool. Transaction &quot;Stand=
ardness&quot; rules and mempool restrictions such<br>as &quot;too-long-memp=
ool-chain&quot; are both examples of policy.<br><br>### Package Relay and P=
ackage Mempool Accept<br><br>In transaction relay, we currently consider tr=
ansactions one at a time for<br>submission to the mempool. This creates a l=
imitation in the node&#39;s ability to<br>determine which transactions have=
 the highest feerates, since we cannot take<br>into account descendants (i.=
e. cannot use CPFP) until all the transactions are<br>in the mempool. Simil=
arly, we cannot use a transaction&#39;s descendants when<br>considering it =
for RBF. When an individual transaction does not meet the mempool<br>minimu=
m feerate and the user isn&#39;t able to create a replacement transaction<b=
r>directly, it will not be accepted by mempools.<br><br>This limitation pre=
sents a security issue for applications and users relying on<br>time-sensit=
ive transactions. For example, Lightning and other protocols create<br>UTXO=
s with multiple spending paths, where one counterparty&#39;s spending path =
opens<br>up after a timelock, and users are protected from cheating scenari=
os as long as<br>they redeem on-chain in time. A key security assumption is=
 that all parties&#39;<br>transactions will propagate and confirm in a time=
ly manner. This assumption can<br>be broken if fee-bumping does not work as=
 intended.<br><br>The end goal for Package Relay is to consider multiple tr=
ansactions at the same<br>time, e.g. a transaction with its high-fee child.=
 This may help us better<br>determine whether transactions should be accept=
ed to our mempool, especially if<br>they don&#39;t meet fee requirements in=
dividually or are better RBF candidates as a<br>package. A combination of c=
hanges to mempool validation logic, policy, and<br>transaction relay allows=
 us to better propagate the transactions with the<br>highest package feerat=
es to miners, and makes fee-bumping tools more powerful<br>for users.<br><b=
r>The &quot;relay&quot; part of Package Relay suggests P2P messaging change=
s, but a large<br>part of the changes are in the mempool&#39;s package vali=
dation logic. We call this<br>*Package Mempool Accept*.<br><br>### Previous=
 Work<br><br>* Given that mempool validation is DoS-sensitive and complex, =
it would be<br>=C2=A0 dangerous to haphazardly tack on package validation l=
ogic. Many efforts have<br>been made to make mempool validation less opaque=
 (see [#16400][4], [#21062][5],<br>[#22675][6], [#22796][7]).<br>* [#20833]=
[8] Added basic capabilities for package validation, test accepts only<br>=
=C2=A0 (no submission to mempool).<br>* [#21800][9] Implemented package anc=
estor/descendant limit checks for arbitrary<br>=C2=A0 packages. Still test =
accepts only.<br>* Previous package relay proposals (see [#16401][10], [#19=
621][11]).<br><br>### Existing Package Rules<br><br>These are in master as =
introduced in [#20833][8] and [#21800][9]. I&#39;ll consider<br>them as &qu=
ot;given&quot; in the rest of this document, though they can be changed, si=
nce<br>package validation is test-accept only right now.<br><br>1. A packag=
e cannot exceed `MAX_PACKAGE_COUNT=3D25` count and<br>`MAX_PACKAGE_SIZE=3D1=
01KvB` total size [8]<br><br>=C2=A0 =C2=A0*Rationale*: This is already enfo=
rced as mempool ancestor/descendant limits.<br>Presumably, transactions in =
a package are all related, so exceeding this limit<br>would mean that the p=
ackage can either be split up or it wouldn&#39;t pass this<br>mempool polic=
y.<br><br>2. Packages must be topologically sorted: if any dependencies exi=
st between<br>transactions, parents must appear somewhere before children. =
[8]<br><br>3. A package cannot have conflicting transactions, i.e. none of =
them can spend<br><div>the same inputs. This also means there cannot be dup=
licate transactions. [8]</div><div><br></div>4. When packages are evaluated=
 against ancestor/descendant limits in a test<br>accept, the union of all o=
f their descendants and ancestors is considered. This<br>is essentially a &=
quot;worst case&quot; heuristic where every transaction in the package<br>i=
s treated as each other&#39;s ancestor and descendant.  [8]<br>Packages for=
 which ancestor/descendant limits are accurately captured by this<br><div>h=
euristic: [19]</div><br>There are also limitations such as the fact that CP=
FP carve out is not applied<br>to package transactions. #20833 also disable=
s RBF in package validation; this<br>proposal overrides that to allow packa=
ges to use RBF.<br><br>## Proposed Changes<br><br>The next step in the Pack=
age Mempool Accept project is to implement submission<br>to mempool, initia=
lly through RPC only. This allows us to test the submission<br>logic before=
 exposing it on P2P.<br><br>### Summary<br><br>- Packages may contain alrea=
dy-in-mempool transactions.<br>- Packages are 2 generations, Multi-Parent-1=
-Child.<br>- Fee-related checks use the package feerate. This means that wa=
llets can<br>create a package that utilizes CPFP.<br>- Parents are allowed =
to RBF mempool transactions with a set of rules similar<br>=C2=A0 to BIP125=
. This enables a combination of CPFP and RBF, where a<br>transaction&#39;s =
descendant fees pay for replacing mempool conflicts.<br><br>There is a draf=
t implementation in [#22290][1]. It is WIP, but feedback is<br>always welco=
me.<br><br>### Details<br><br>#### Packages May Contain Already-in-Mempool =
Transactions<br><br>A package may contain transactions that are already in =
the mempool. We remove<br>(&quot;deduplicate&quot;) those transactions from=
 the package for the purposes of package<br>mempool acceptance. If a packag=
e is empty after deduplication, we do nothing.<br><br>*Rationale*: Mempools=
 vary across the network. It&#39;s possible for a parent to be<br>accepted =
to the mempool of a peer on its own due to differences in policy and<br>fee=
 market fluctuations. We should not reject or penalize the entire package f=
or<br>an individual transaction as that could be a censorship vector.<br><b=
r>#### Packages Are Multi-Parent-1-Child<br><br>Only packages of a specific=
 topology are permitted. Namely, a package is exactly<br>1 child with all o=
f its unconfirmed parents. After deduplication, the package<br>may be exact=
ly the same, empty, 1 child, 1 child with just some of its<br>unconfirmed p=
arents, etc. Note that it&#39;s possible for the parents to be indirect<br>=
descendants/ancestors of one another, or for parent and child to share a pa=
rent,<br>so we cannot make any other topology assumptions.<br><br>*Rational=
e*: This allows for fee-bumping by CPFP. Allowing multiple parents<br>makes=
 it possible to fee-bump a batch of transactions. Restricting packages to a=
<br>defined topology is also easier to reason about and simplifies the vali=
dation<br>logic greatly. Multi-parent-1-child allows us to think of the pac=
kage as one big<br>transaction, where:<br><br>- Inputs =3D all the inputs o=
f parents + inputs of the child that come from<br>=C2=A0 confirmed UTXOs<br=
>- Outputs =3D all the outputs of the child + all outputs of the parents th=
at<br>=C2=A0 aren&#39;t spent by other transactions in the package<br><br>E=
xamples of packages that follow this rule (variations of example A show som=
e<br>possibilities after deduplication): ![image][15]<br><br>#### Fee-Relat=
ed Checks Use Package Feerate<br><br>Package Feerate =3D the total modified=
 fees divided by the total virtual size of<br>all transactions in the packa=
ge.<br><br>To meet the two feerate requirements of a mempool, i.e., the pre=
-configured<br>minimum relay feerate (`minRelayTxFee`) and dynamic mempool =
minimum feerate, the<br>total package feerate is used instead of the indivi=
dual feerate. The individual<br>transactions are allowed to be below feerat=
e requirements if the package meets<br>the feerate requirements. For exampl=
e, the parent(s) in the package can have 0<br>fees but be paid for by the c=
hild.<br><br>*Rationale*: This can be thought of as &quot;CPFP within a pac=
kage,&quot; solving the<br>issue of a parent not meeting minimum fees on it=
s own. This allows L2<br>applications to adjust their fees at broadcast tim=
e instead of overshooting or<br>risking getting stuck/pinned.<br><br>We use=
 the package feerate of the package *after deduplication*.<br><br>*Rational=
e*: =C2=A0It would be incorrect to use the fees of transactions that are<br=
>already in the mempool, as we do not want a transaction&#39;s fees to be<b=
r>double-counted for both its individual RBF and package RBF.<br><br>Exampl=
es F and G [14] show the same package, but P1 is submitted individually bef=
ore<br>the package in example G. In example F, we can see that the 300vB pa=
ckage pays<br>an additional 200sat in fees, which is not enough to pay for =
its own bandwidth<br>(BIP125#4). In example G, we can see that P1 pays enou=
gh to replace M1, but<br>using P1&#39;s fees again during package submissio=
n would make it look like a 300sat<br>increase for a 200vB package. Even in=
cluding its fees and size would not be<br>sufficient in this example, since=
 the 300sat looks like enough for the 300vB<br>package. The calculcation af=
ter deduplication is 100sat increase for a package<br>of size 200vB, which =
correctly fails BIP125#4. Assume all transactions have a<br>size of 100vB.<=
br><br>#### Package RBF<br><br>If a package meets feerate requirements as a=
 package, the parents in the<br>transaction are allowed to replace-by-fee m=
empool transactions. The child cannot<br>replace mempool transactions. Mult=
iple transactions can replace the same<br>transaction, but in order to be v=
alid, none of the transactions can try to<br>replace an ancestor of another=
 transaction in the same package (which would thus<br>make its inputs unava=
ilable).<br><br>*Rationale*: Even if we are using package feerate, a packag=
e will not propagate<br>as intended if RBF still requires each individual t=
ransaction to meet the<br>feerate requirements.<br><br>We use a set of rule=
s slightly modified from BIP125 as follows:<br><br>##### Signaling (Rule #1=
)<br><br>All mempool transactions to be replaced must signal replaceability=
.<br><br>*Rationale*: Package RBF signaling logic should be the same for pa=
ckage RBF and<br>single transaction acceptance. This would be updated if si=
ngle transaction<br>validation moves to full RBF.<br><br>##### New Unconfir=
med Inputs (Rule #2)<br><br>A package may include new unconfirmed inputs, b=
ut the ancestor feerate of the<br>child must be at least as high as the anc=
estor feerates of every transaction<br>being replaced. This is contrary to =
BIP125#2, which states &quot;The replacement<br>transaction may only includ=
e an unconfirmed input if that input was included in<br>one of the original=
 transactions. (An unconfirmed input spends an output from a<br>currently-u=
nconfirmed transaction.)&quot;<br><br>*Rationale*: The purpose of BIP125#2 =
is to ensure that the replacement<br>transaction has a higher ancestor scor=
e than the original transaction(s) (see<br>[comment][13]). Example H [16] s=
hows how adding a new unconfirmed input can lower the<br>ancestor score of =
the replacement transaction. P1 is trying to replace M1, and<br>spends an u=
nconfirmed output of M2. P1 pays 800sat, M1 pays 600sat, and M2 pays<br>100=
sat. Assume all transactions have a size of 100vB. While, in isolation, P1<=
br>looks like a better mining candidate than M1, it must be mined with M2, =
so its<br>ancestor feerate is actually 4.5sat/vB.=C2=A0 This is lower than =
M1&#39;s ancestor<br>feerate, which is 6sat/vB.<br><br>In package RBF, the =
rule analogous to BIP125#2 would be &quot;none of the<br>transactions in th=
e package can spend new unconfirmed inputs.&quot; Example J [17] shows<br>w=
hy, if any of the package transactions have ancestors, package feerate is n=
o<br>longer accurate. Even though M2 and M3 are not ancestors of P1 (which =
is the<br>replacement transaction in an RBF), we&#39;re actually interested=
 in the entire<br>package. A miner should mine M1 which is 5sat/vB instead =
of M2, M3, P1, P2, and<br>P3, which is only 4sat/vB. The Package RBF rule c=
annot be loosened to only allow<br>the child to have new unconfirmed inputs=
, either, because it can still cause us<br>to overestimate the package&#39;=
s ancestor score.<br><br>However, enforcing a rule analogous to BIP125#2 wo=
uld not only make Package RBF<br>less useful, but would also break Package =
RBF for packages with parents already<br>in the mempool: if a package paren=
t has already been submitted, it would look<br>like the child is spending a=
 &quot;new&quot; unconfirmed input. In example K [18], we&#39;re<br>looking=
 to replace M1 with the entire package including P1, P2, and P3. We must<br=
>consider the case where one of the parents is already in the mempool (in t=
his<br>case, P2), which means we must allow P3 to have new unconfirmed inpu=
ts. However,<br>M2 lowers the ancestor score of P3 to 4.3sat/vB, so we shou=
ld not replace M1<br>with this package.<br><br>Thus, the package RBF rule r=
egarding new unconfirmed inputs is less strict than<br>BIP125#2. However, w=
e still achieve the same goal of requiring the replacement<br>transactions =
to have a ancestor score at least as high as the original ones. As<br>a res=
ult, the entire package is required to be a higher feerate mining candidate=
<br>than each of the replaced transactions.<br><br>Another note: the [comme=
nt][13] above the BIP125#2 code in the original RBF<br>implementation sugge=
sts that the rule was intended to be temporary.<br><br>##### Absolute Fee (=
Rule #3)<br><br>The package must increase the absolute fee of the mempool, =
i.e. the total fees<br>of the package must be higher than the absolute fees=
 of the mempool transactions<br>it replaces. Combined with the CPFP rule ab=
ove, this differs from BIP125 Rule #3<br>- an individual transaction in the=
 package may have lower fees than the<br>=C2=A0 transaction(s) it is replac=
ing. In fact, it may have 0 fees, and the child<br>pays for RBF.<br><br>###=
## Feerate (Rule #4)<br><br>The package must pay for its own bandwidth; the=
 package feerate must be higher<br>than the replaced transactions by at lea=
st minimum relay feerate<br>(`incrementalRelayFee`). Combined with the CPFP=
 rule above, this differs from<br>BIP125 Rule #4 - an individual transactio=
n in the package can have a lower<br>feerate than the transaction(s) it is =
replacing. In fact, it may have 0 fees,<br>and the child pays for RBF.<br><=
br>##### Total Number of Replaced Transactions (Rule #5)<br><br>The package=
 cannot replace more than 100 mempool transactions. This is identical<br>to=
 BIP125 Rule #5.<br><br>### Expected FAQs<br><br>1. Is it possible for only=
 some of the package to make it into the mempool?<br><br>=C2=A0 =C2=A0Yes, =
it is. However, since we evict transactions from the mempool by<br>descenda=
nt score and the package child is supposed to be sponsoring the fees of<br>=
its parents, the most common scenario would be all-or-nothing. This is<br>i=
ncentive-compatible. In fact, to be conservative, package validation should=
<br>begin by trying to submit all of the transactions individually, and onl=
y use the<br>package mempool acceptance logic if the parents fail due to lo=
w feerate.<br><br>2. Should we allow packages to contain already-confirmed =
transactions?<br><br>=C2=A0 =C2=A0 No, for practical reasons. In mempool va=
lidation, we actually aren&#39;t able to<br>tell with 100% confidence if we=
 are looking at a transaction that has already<br>confirmed, because we loo=
k up inputs using a UTXO set. If we have historical<br>block data, it&#39;s=
 possible to look for it, but this is inefficient, not always<br>possible f=
or pruning nodes, and unnecessary because we&#39;re not going to do<br>anyt=
hing with the transaction anyway. As such, we already have the expectation<=
br>that transaction relay is somewhat &quot;stateful&quot; i.e. nobody shou=
ld be relaying<br>transactions that have already been confirmed. Similarly,=
 we shouldn&#39;t be<br>relaying packages that contain already-confirmed tr=
ansactions.<br><br>[1]: <a href=3D"https://github.com/bitcoin/bitcoin/pull/=
22290" target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/22290</a><=
br>[2]: <a href=3D"https://github.com/bitcoin/bips/blob/1f0b563738199ca60d3=
2b4ba779797fc97d040fe/bip-0141.mediawiki#transaction-size-calculations" tar=
get=3D"_blank">https://github.com/bitcoin/bips/blob/1f0b563738199ca60d32b4b=
a779797fc97d040fe/bip-0141.mediawiki#transaction-size-calculations</a><br>[=
3]: <a href=3D"https://github.com/bitcoin/bitcoin/blob/94f83534e4b771944af7=
d9ed0f40746f392eb75e/src/policy/policy.cpp#L282" target=3D"_blank">https://=
github.com/bitcoin/bitcoin/blob/94f83534e4b771944af7d9ed0f40746f392eb75e/sr=
c/policy/policy.cpp#L282</a><br>[4]: <a href=3D"https://github.com/bitcoin/=
bitcoin/pull/16400" target=3D"_blank">https://github.com/bitcoin/bitcoin/pu=
ll/16400</a><br>[5]: <a href=3D"https://github.com/bitcoin/bitcoin/pull/210=
62" target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/21062</a><br>=
[6]: <a href=3D"https://github.com/bitcoin/bitcoin/pull/22675" target=3D"_b=
lank">https://github.com/bitcoin/bitcoin/pull/22675</a><br>[7]: <a href=3D"=
https://github.com/bitcoin/bitcoin/pull/22796" target=3D"_blank">https://gi=
thub.com/bitcoin/bitcoin/pull/22796</a><br>[8]: <a href=3D"https://github.c=
om/bitcoin/bitcoin/pull/20833" target=3D"_blank">https://github.com/bitcoin=
/bitcoin/pull/20833</a><br>[9]: <a href=3D"https://github.com/bitcoin/bitco=
in/pull/21800" target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/21=
800</a><br>[10]: <a href=3D"https://github.com/bitcoin/bitcoin/pull/16401" =
target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/16401</a><br>[11]=
: <a href=3D"https://github.com/bitcoin/bitcoin/pull/19621" target=3D"_blan=
k">https://github.com/bitcoin/bitcoin/pull/19621</a><br>[12]: <a href=3D"ht=
tps://github.com/bitcoin/bips/blob/master/bip-0125.mediawiki" target=3D"_bl=
ank">https://github.com/bitcoin/bips/blob/master/bip-0125.mediawiki</a><br>=
[13]: <a href=3D"https://github.com/bitcoin/bitcoin/pull/6871/files#diff-34=
d21af3c614ea3cee120df276c9c4ae95053830d7f1d3deaf009a4625409ad2R1101-R1104" =
target=3D"_blank">https://github.com/bitcoin/bitcoin/pull/6871/files#diff-3=
4d21af3c614ea3cee120df276c9c4ae95053830d7f1d3deaf009a4625409ad2R1101-R1104<=
/a><br>[14]: <a href=3D"https://user-images.githubusercontent.com/25183001/=
133567078-075a971c-0619-4339-9168-b41fd2b90c28.png" target=3D"_blank">https=
://user-images.githubusercontent.com/25183001/133567078-075a971c-0619-4339-=
9168-b41fd2b90c28.png</a><br>[15]: <a href=3D"https://user-images.githubuse=
rcontent.com/25183001/132856734-fc17da75-f875-44bb-b954-cb7a1725cc0d.png" t=
arget=3D"_blank">https://user-images.githubusercontent.com/25183001/1328567=
34-fc17da75-f875-44bb-b954-cb7a1725cc0d.png</a><br>[16]: <a href=3D"https:/=
/user-images.githubusercontent.com/25183001/133567347-a3e2e4a8-ae9c-49f8-ab=
b9-81e8e0aba224.png" target=3D"_blank">https://user-images.githubuserconten=
t.com/25183001/133567347-a3e2e4a8-ae9c-49f8-abb9-81e8e0aba224.png</a><br>[1=
7]: <a href=3D"https://user-images.githubusercontent.com/25183001/133567370=
-21566d0e-36c8-4831-b1a8-706634540af3.png" target=3D"_blank">https://user-i=
mages.githubusercontent.com/25183001/133567370-21566d0e-36c8-4831-b1a8-7066=
34540af3.png</a><br>[18]: <a href=3D"https://user-images.githubusercontent.=
com/25183001/133567444-bfff1142-439f-4547-800a-2ba2b0242bcb.png" target=3D"=
_blank">https://user-images.githubusercontent.com/25183001/133567444-bfff11=
42-439f-4547-800a-2ba2b0242bcb.png</a><br>[19]: <a href=3D"https://user-ima=
ges.githubusercontent.com/25183001/133456219-0bb447cb-dcb4-4a31-b9c1-7d8620=
5b68bc.png" target=3D"_blank">https://user-images.githubusercontent.com/251=
83001/133456219-0bb447cb-dcb4-4a31-b9c1-7d86205b68bc.png</a><br>[20]: <a hr=
ef=3D"https://user-images.githubusercontent.com/25183001/132857787-7b7c6f56=
-af96-44c8-8d78-983719888c19.png" target=3D"_blank">https://user-images.git=
hubusercontent.com/25183001/132857787-7b7c6f56-af96-44c8-8d78-983719888c19.=
png</a><br></div>
_______________________________________________<br>
bitcoin-dev mailing list<br>
<a href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" target=3D"_blank">=
bitcoin-dev@lists.linuxfoundation.org</a><br>
<a href=3D"https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev" =
rel=3D"noreferrer" target=3D"_blank">https://lists.linuxfoundation.org/mail=
man/listinfo/bitcoin-dev</a><br>
</blockquote></div>
</blockquote></div>
</blockquote></div>
</blockquote></div>
</blockquote></div>
_______________________________________________<br>
bitcoin-dev mailing list<br>
<a href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" target=3D"_blank">=
bitcoin-dev@lists.linuxfoundation.org</a><br>
<a href=3D"https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev" =
rel=3D"noreferrer" target=3D"_blank">https://lists.linuxfoundation.org/mail=
man/listinfo/bitcoin-dev</a><br>
</blockquote></div>

--000000000000a309eb05ccf4dc02--