Date: Tue, 15 Oct 2019 07:20:07 +0000
To: Braydon Fuller <braydon@purse.io>,
	Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>
From: =?UTF-8?Q?Joachim_Str=C3=B6mbergson?= <joachimstr@protonmail.com>
Reply-To: =?UTF-8?Q?Joachim_Str=C3=B6mbergson?= <joachimstr@protonmail.com>
Message-ID: <yKuDn4LKCA9-olIsgnJO7VqN3RE9BgZWEzLu6-k4wG30iHYDIJa6wjTBoYXU0zWCRyNCEvBb4xQw7qp0BDGBCR4zccI6kJ_4AnlsB9vc9L4=@protonmail.com>
In-Reply-To: <93649df9-27ab-abaf-00f3-da6c528344cc@purse.io>
References: <42cd5ffd-63e8-b738-c4ea-13d0699b1268@purse.io>
	<CAE-z3OV_LL+Jww3e=gO6t=02VW7m9PK+8EaYoEVLy9NKNMiSaQ@mail.gmail.com>
	<e9c5e519-ea8a-f0e2-d8fb-c955b5c2de40@purse.io>
	<CAE-z3OXyTc0aoJJVNLS5MReE7+Nhckyrjf22+yCSjXF8=bNbXQ@mail.gmail.com>
	<H_Yq1W3SffFweLPPXiUiA4EeU2yU7c8LVcqw5AbajovWTWMt5hKQARKglEQwCjPpXvjiBfvmTnaXJwivkGkT8BDha8k303DNbFB-ECes0d4=@protonmail.com>
	<93649df9-27ab-abaf-00f3-da6c528344cc@purse.io>
Feedback-ID: rtGq1wInl4cYyZOA2iZwaHP-4FBFY67Qt3DcVBMZh8YR1tV-3hijnv7SxpdDwGlNdSPgHEykKLn6PcHDKa0D8A==:Ext:ProtonMail
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [bitcoin-dev] Chain width expansion
Precedence: list


> > [...] First you provide proof of your best block height via coinbase [.=
..]
>
> So I don't think you can use the height in the coinbase for that
> purpose, as it's not possible to validate it without the previous
> headers. That's common for more than just the height.

You can fake it of course, but it means you are spending money for PoW and =
I can't see viable attack here. You can commit to either lower height than =
actual or higher height, if you are malicious. If it is lower, then your ch=
ain will look weaker with some heuristic based on PoW of the tip and the ch=
ain length. So you are not going to do that. It thus seem it only makes sen=
se to commit to higher than actual height. OK, this could convince me to be=
 interested in your chain over others. So let's say the real chain is 600k =
blocks, you claim 1m blocks, and you prove 1m height block to me. So I can =
ask you about

1) 2000 headers from the tip down
2) AND another proof of height for the block at 1m-2000.

To be able to provide that you need to have such a chain and you can reuse =
any real subchain from mainnet. It's still possible that you can deliver th=
at, but not for high difficulty.

Now if time warp attack is blocked, you will have pretty hard time to creat=
e such a chain that would look strongest in cumulative work. I don't have a=
ctual numbers though, so I can't say this would mitigate everything.


> > [...] to generate much longer chain with superslow timestamp increase (=
~5 blocks in 1 second) without increasing difficulty (i.e. staying at min. =
diff.). [...]
>
> In that case, it would take about 7 minutes of block time seconds for
> the next retarget period, every 2016 blocks, and the difficulty would
> adjust. The difficulty would adjust in that case as if 2 weeks of blocks
> had been mined in 7 minutes. For the difficulty to remain the same the
> time between blocks needs to be 10 minutes.

This calculation does not apply under time warp attack. You can fake timest=
amps of all blocks except for those relevant to the retarget calculation. T=
hose are only the first and the last block in the 2016 block window.