MIME-Version: 1.0
In-Reply-To: <CAPkFh0vNECi1OmBwki+8NNAQbe6EG2FEE4RR5z=kYVLLDFHUXg@mail.gmail.com>
References: <20151219184240.GB12893@muck>
	<CAAcC9yvh2ma2dFhNDEKs7vfXyQF9L+T0YtRvOsJ15AbfVti=cw@mail.gmail.com>
	<219f125cee6ca68fd27016642e38fdf1@xbt.hk>
	<CAAcC9ys_t7X0WpQ8W3577M8GLiA5sPV2F1BJ9qZbnMkE-1j3+Q@mail.gmail.com>
	<aff8da46a69bdd7ef92ca87725866a5c@xbt.hk>
	<CAPkFh0vNECi1OmBwki+8NNAQbe6EG2FEE4RR5z=kYVLLDFHUXg@mail.gmail.com>
Date: Sun, 20 Dec 2015 11:38:34 +0000
Message-ID: <CAE-z3OU=18VuV+9U9meg5fRxQt3MZnAnQ2jPN5QBNk+ZtSoJXw@mail.gmail.com>
From: Tier Nolan <tier.nolan@gmail.com>
Cc: Bitcoin Dev <bitcoin-dev@lists.linuxfoundation.org>
Content-Type: multipart/alternative; boundary=089e010d931a022c8e052752d086
Subject: Re: [bitcoin-dev] We need to fix the block withholding attack
Precedence: list

--089e010d931a022c8e052752d086
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

On Sun, Dec 20, 2015 at 5:12 AM, Emin G=C3=BCn Sirer <
bitcoin-dev@lists.linuxfoundation.org> wrote:

>  An attacker pool (A) can take a certain portion of its hashpower,
> use it to mine on behalf of victim pool (B), furnish partial proofs of wo=
rk
> to B, but discard any full blocks it discovers.
>

I wonder if part of the problem here is that there is no pool identity
linked to mining pools.

If the mining protocols were altered so that miners had to indicate their
identity, then a pool couldn't forward hashing power to their victim.

If the various mining protocols were updated, they could allow checking
that the work has the domain name of the pool included.  Pools would have
to include their domain name in the block header.

A pool which provides this service is publicly saying that they will not
use the block withholding attack.  Any two pools which are doing it cannot
attack each other (since they have different domain names).  This creates
an incentive for pools to start supporting the feature.

Owners of hashing power also have an incentive to operate with pools which
offer this identity.  It means that they can ensure that they get a payout
from any blocks found.

Hosted mining is weaker, but even then, it is possible for mining hosts to
provide proof that they performed mining.  This proof would include the
identity of the mining pool.  Even if the pool was run by the host, it
would still need to have the name embedded.

Mining hosts might be able to figure out which of their customers actually
check the identity info, and then they could redirect the mining power of
those who generally don't check.  If customers randomly ask for all of the
hashing power, right back to when they joined, then this becomes expensive.

Mining power directly owned by the pool is also immune to this effect.

--089e010d931a022c8e052752d086
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_extra"><div class=3D"gmail_quote">On S=
un, Dec 20, 2015 at 5:12 AM, Emin G=C3=BCn Sirer <span dir=3D"ltr">&lt;<a h=
ref=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" target=3D"_blank">bitc=
oin-dev@lists.linuxfoundation.org</a>&gt;</span> wrote:<br><blockquote clas=
s=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;pad=
ding-left:1ex"><div dir=3D"ltr">=C2=A0An attacker pool (A) can take a certa=
in portion of its hashpower,<div>use it to mine on behalf of victim pool (B=
), furnish partial proofs of work</div><div>to B, but discard any full bloc=
ks it discovers.</div></div></blockquote><div><br></div><div>I wonder if pa=
rt of the problem here is that there is no pool identity linked to mining p=
ools.<br><br></div><div>If the mining protocols were altered so that miners=
 had to indicate their identity, then a pool couldn&#39;t forward hashing p=
ower to their victim.=C2=A0 <br><br></div><div>If the various mining protoc=
ols were updated, they could allow checking that the work has the domain na=
me of the pool included.=C2=A0 Pools would have to include their domain nam=
e in the block header.<br><br></div><div>A pool which provides this service=
 is publicly saying that they will not use the block withholding attack.=C2=
=A0 Any two pools which are doing it cannot attack each other (since they h=
ave different domain names).=C2=A0 This creates an incentive for pools to s=
tart supporting the feature.<br><br></div><div>Owners of hashing power also=
 have an incentive to operate with pools which offer this identity.=C2=A0 I=
t means that they can ensure that they get a payout from any blocks found.<=
br><br></div><div>Hosted mining is weaker, but even then, it is possible fo=
r mining hosts to provide proof that they performed mining.=C2=A0 This proo=
f would include the identity of the mining pool.=C2=A0 Even if the pool was=
 run by the host, it would still need to have the name embedded.=C2=A0 <br>=
<br></div><div>Mining hosts might be able to figure out which of their cust=
omers actually check the identity info, and then they could redirect the mi=
ning power of those who generally don&#39;t check.=C2=A0 If customers rando=
mly ask for all of the hashing power, right back to when they joined, then =
this becomes expensive.<br><br></div><div>Mining power directly owned by th=
e pool is also immune to this effect.<br></div></div></div></div>

--089e010d931a022c8e052752d086--