MIME-Version: 1.0
In-Reply-To: <CACDYSUROutAMV7C8pUXz0PMvH5awkw-XUtce7BxTtZMD_yUm5A@mail.gmail.com>
References: <CAJR7vkpRhNsQsem-nFkeubX04xx1y7aHwCENfg0d1266oOsXMw@mail.gmail.com>
	<Cwhn7YzwaDUZtOygDAgrU1UXjRPG-EiH3Fyz2c95gqOpNnNbiYL1NvhS28yK5wLJCnIqDaBrM6c574dY-O6_-bRjLIFmDe2NCxIuyV1w2dw=@protonmail.com>
	<CAJR7vkoq8Y_-fbdxN=--gF5wrGByr5oODc4FkTaCEvDSuP0whQ@mail.gmail.com>
	<CACDYSUROutAMV7C8pUXz0PMvH5awkw-XUtce7BxTtZMD_yUm5A@mail.gmail.com>
From: Jimmy Song <jaejoon@gmail.com>
Date: Sat, 8 Apr 2017 09:35:30 -0500
Message-ID: <CAJR7vkrn-oFium3wFgcOdqNPuYq+rW2DqyOnkDaCTHabO0y3Xg@mail.gmail.com>
To: Pavel Moravec <pavel.moravec@braiins.cz>
Content-Type: multipart/alternative; boundary=001a1148e60a7622b0054ca8a77f
Cc: Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] A Small Modification to Segwit
Precedence: list

--001a1148e60a7622b0054ca8a77f
Content-Type: text/plain; charset=UTF-8

Pavel,

Until all miners update (firmware or hardware), the change encourages
> large difference in mining efficiency. And IMO it gives another
> advantage to large mining operations in general.
>

Certainly, there would have to be changes for stratum, pool software, etc.
But the monetary incentives align to all the changes needed.

Remember, overt ASICBoost can get something like a 12.5% efficiency boost
from toggling a single bit in the version (equivalent to 2 colliding work
items), 18.5% from 2 bits (equivalent to 4 colliding work items), 23.4%
from 4 bits (see https://arxiv.org/ftp/arxiv/papers/1604/1604.00575.pdf).
In lieu of an explicit allowance of overt ASICBoost, the monetary
incentives lead to odd BIP9 signaling, especially if 4 or more proposals
signal at once. There really isn't a practical way to block overt ASICBoost
without forcing the version bits to be some value.

In other words, the question isn't about allowing/disallowing ASICBoost at
this point. The question is whether we want ASICBoost open or hidden.


> You make a strong assumption that the new optimization is not
> compatible with overt ASICBoost. If it is compatible, ASICBoost
> doesn't help you with "defending against" the new optimization at all.
> And it can be the case that the new optimization is based on ASICBoost
> so you can make the situation "worse" by allowing it.
>

This would only be the case if overt ASICBoost were not possible at all. It
is currently possible to use overt ASICBoost, so optimizations based on
overt ASICBoost would also be possible unless something were done to
actively block it.

> Certainly, if only one company made use of the extra nonce space, they
> would have an advantage.
>
> Can you explain why the reality should be significantly different? In
> sufficiently near future.


Market incentives, I would imagine. How quickly that would be is not
something I'm qualified to answer.


> We don't have to deal with any such theoretical situation now. You
> proposal goes in opposite direction, by adding support for patented
> algorithm. I don't know myself what the possible legal implications
> are (maybe only for a subset of miners) so I consider it as an
> unnecessary risk. At least before some conclusive legal analysis says
> differently.
>

I'm not adding support as much as explicitly allowing what's implicitly
allowed. Whatever risks you imagine for this proposal exist on the network
currently, with unmodified BIP-141 and with modified BIP-141. The
difference in adding the modification is that overt ASICBoost is explicitly
allowed in the modified BIP-141 as to not hide it.

Jimmy

--001a1148e60a7622b0054ca8a77f
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"><div=
>Pavel,</div><div><br></div><blockquote class=3D"gmail_quote" style=3D"marg=
in:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1e=
x">
Until all miners update (firmware or hardware), the change encourages<br>
large difference in mining efficiency. And IMO it gives another<br>
advantage to large mining operations in general.<span class=3D"gmail-"><br>=
</span></blockquote><div><br></div><div>Certainly, there would have to be c=
hanges for stratum, pool software, etc. But the monetary incentives align t=
o all the changes needed.=C2=A0</div><div><br></div><div>Remember, overt AS=
ICBoost can get something like a 12.5% efficiency boost from toggling a sin=
gle bit in the version (equivalent to 2 colliding work items), 18.5% from 2=
 bits (equivalent to 4 colliding work items), 23.4% from 4 bits (see <a hre=
f=3D"https://arxiv.org/ftp/arxiv/papers/1604/1604.00575.pdf">https://arxiv.=
org/ftp/arxiv/papers/1604/1604.00575.pdf</a>). In lieu of an explicit allow=
ance of overt ASICBoost, the monetary incentives lead to odd BIP9 signaling=
, especially if 4 or more proposals signal at once. There really isn&#39;t =
a practical way to block overt ASICBoost without forcing the version bits t=
o be some value.</div><div><br></div><div>In other words, the question isn&=
#39;t about allowing/disallowing ASICBoost at this point. The question is w=
hether we want ASICBoost open or hidden.</div><div>=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">You make a strong assumption that=
 the new optimization is not<br>
compatible with overt ASICBoost. If it is compatible, ASICBoost<br>
doesn&#39;t help you with &quot;defending against&quot; the new optimizatio=
n at all.<br>
And it can be the case that the new optimization is based on ASICBoost<br>
so you can make the situation &quot;worse&quot; by allowing it.<br></blockq=
uote><div><br></div><div>This would only be the case if overt ASICBoost wer=
e not possible at all. It is currently possible to use overt ASICBoost, so =
optimizations based on overt ASICBoost would also be possible unless someth=
ing were done to actively block it.</div><div><br></div><blockquote class=
=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rg=
b(204,204,204);padding-left:1ex"><span class=3D"gmail-">
&gt; Certainly, if only one company made use of the extra nonce space, they=
 would have an advantage.<br>
<br>
</span>Can you explain why the reality should be significantly different? I=
n<br>
sufficiently near future.</blockquote><div><br></div><div>Market incentives=
, I would imagine. How quickly that would be is not something I&#39;m quali=
fied to answer.<br></div><div>=C2=A0</div><blockquote class=3D"gmail_quote"=
 style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);p=
adding-left:1ex">We don&#39;t have to deal with any such theoretical situat=
ion now. You<br>
proposal goes in opposite direction, by adding support for patented<br>
algorithm. I don&#39;t know myself what the possible legal implications<br>
are (maybe only for a subset of miners) so I consider it as an<br>
unnecessary risk. At least before some conclusive legal analysis says<br>
differently.<br>
</blockquote></div><br></div><div class=3D"gmail_extra">I&#39;m not adding =
support as much as explicitly allowing what&#39;s implicitly allowed. Whate=
ver risks you imagine for this proposal exist on the network currently, wit=
h unmodified BIP-141 and with modified BIP-141. The difference in adding th=
e modification is that overt ASICBoost is explicitly allowed in the modifie=
d BIP-141 as to not hide it.</div><div class=3D"gmail_extra"><br></div><div=
 class=3D"gmail_extra">Jimmy</div></div>

--001a1148e60a7622b0054ca8a77f--