Date: Mon, 30 Mar 2020 01:25:36 +0000
To: Ruben Somsen <rsomsen@gmail.com>
From: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Reply-To: ZmnSCPxj <ZmnSCPxj@protonmail.com>
Message-ID: <70epI8yeOu69uXQfxbyWEfrH7hYLzsx9CDIgA9gL_GlaqDEmshtP4Ogf6Dl7GH408GTPDveir1MKy1euEcPbOhJEtzjLbV9m506quXhnKOg=@protonmail.com>
In-Reply-To: <CAPv7Tjb5a5RbXH802m8qHoUKw-5rZV6nOw01z9+hx4yTJYV3Cw@mail.gmail.com>
References: <CAJvkSseW9OZ50yQiS7e0zt9tQt4v9aoikgGs_54_kMN-ORkQgw@mail.gmail.com>
 <20200327171017.GM28113@mcelrath.org>
 <6SFA-3YOsuEl4D3N5eA5G6q1B6ROWHmNefjCC5cRPpHg6iu9PVdG21PKjV28IMYXY_tOcmrSB60tnQRgm4pcHXB_MxOPaa9zZIbBeo0aHS4=@protonmail.com>
 <CAPv7TjYs8j=rKWPVzfFbtznjFQfpKTBGr4AnCXrDcU64Cb8S9Q@mail.gmail.com>
 <CAPv7Tjb5a5RbXH802m8qHoUKw-5rZV6nOw01z9+hx4yTJYV3Cw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Cc: Bitcoin Protocol Discussion <bitcoin-dev@lists.linuxfoundation.org>,
 Tom Trevethan <tom@commerceblock.com>
Subject: Re: [bitcoin-dev] Statechain implementations
Precedence: list

Good morning Ruben,

> Hi ZmnSCPxj,
>
> > the current owner can ask the statechain entity to sign an alternative =
to the first stage, with 0 relative locktime
>
> Unless I am misunderstanding something, this seems to run into the proble=
m that the original first stage transaction is already out there (and its r=
elative timelock started ticking). There is no mechanism ensuring that the =
new tx will have precedence. And even if it did work, I doubt it's cleaner =
than doing a cooperative peg-out that simultaneously happens to peg back in=
, creating a brand new statechain UTXO with no history.


If:

* You are sure the old first stage tx has > 0 relative locktime.
* The replacement tx (which replaces the old first stage) has a 0 relative =
locktime.
  * The replacement tx redirects the funds to a new funding output for a (l=
ogically continuous, onchain new) statechain.

Then the replacement tx, having a smaller relative locktime than the old fi=
rst stage, has precedence.
Indeed, having a *smaller* relative locktime is exactly the mechanism Decke=
r-Wattenhofer uses.

So this is the state, with the kickoff having just been confirmed onchain:


    ***blockchain***
       [funding tx]->[kickoff tx]-+
         _ _ _ _ _ _ _ _ _ _ _ _ _|_ _ _
     offchain                     |
                                  +->[[ 7] stage]->[[ 0] stage]->[[14] stag=
e]-> state outputs

Since the first stage is still "ticking" it is not yet confirmable onchain.

You ask the statechain to create an alternative, 0-relative-locktime, re-fu=
nding tx, and create a new mechanism:

    ***blockchain***
       [funding tx]->[kickoff tx]-+
         _ _ _ _ _ _ _ _ _ _ _ _ _|_ _ _
     offchain                     |
                                  +->[[ 7] stage]->[[ 0] stage]->[[14] stag=
e]-> state outputs
                                 (OR)
                                  |
                                  +->[[ 0] funding tx]->[kickoff tx]->[[14]=
 stage]->[[14] stage]->[[14] stage]->state outputs

Because it has a time advantage, this new re-funding tx has higher priority=
 (and is the same mechanism Decker-Wattenhofer has anyway).

Regards,
ZmnSCPxj