W3C home > Mailing lists > Public > public-interledger@w3.org > January 2016

Re: ILP and signatures

From: Evan Schwartz <evan@ripple.com>
Date: Tue, 26 Jan 2016 15:08:58 -0500
Message-ID: <CAONA2jXetjuuwn4ct78f9hi4K3Bw-mLaDNFF_WfB4suFSbhayA@mail.gmail.com>
To: Melvin Carvalho <melvincarvalho@gmail.com>
Cc: Stefan Thomas <stefan@ripple.com>, Interledger Community Group <public-interledger@w3.org>
>
> What about other kind of "web smart contracts" (for want of a better name)
> ... ie to trigger the release of a payment when something on the web
> happens.  e.g. the ledger operator could publish a receipt of the funds
> transfer to escrow?


We don't specify what type of event or system should produce the
cryptographic condition fulfillment so it definitely could be some kind of
smart contract. Arguably, the most straightforward way for all the ledgers
to check if the predefined event has happened is by looking for a valid
signature that fulfills the condition. The problem with an alternative like
checking a given URI is that the domain owner could respond differently to
different ledgers, causing the payment to be partially executed and for
some parties to lose money.

 Are connector and the escrow agent different?


In general the escrow agent should be the same as the controller of the
*ledger*. In certain cases, such as with banks, the ledger and some
connectors may be operated by the same party but in general they should be
thought of as separate. For something like Bitcoin, the Bitcoin ledger
itself would enforce the conditions of the escrow using its scripting
functionality, while the connector would be a separate, untrusted party
providing the liquidity for the payment.

On Tue, Jan 26, 2016 at 2:41 PM, Melvin Carvalho <melvincarvalho@gmail.com>
wrote:

>
>
> On 26 January 2016 at 19:27, Stefan Thomas <stefan@ripple.com> wrote:
>
>> If you want escrow, the funds' release has to be triggered somehow.
>> Cryptography (hashes or signatures both work) is very convenient for this,
>> but you can in principle use any trusted signal, like a wire with physical
>> security.
>>
>> As far as making ILP a standard I think we'd want to use common
>> cryptographic primitives.
>>
> One more question (apologies if this is a obvious).  In the scenario where
> there is one connector.  Are connector and the escrow agent different?
>
>
>> On Jan 26, 2016 9:03 AM, "Melvin Carvalho" <melvincarvalho@gmail.com>
>> wrote:
>>
>>> Cryptographic signatures are a simple way for ledgers to securely
>>> validate the outcome of the external conditions upon which a transfer is
>>> escrowed. Any one-way function can be used [18]. Using asymmetric
>>> cryptography, the ledger escrows funds pending the presentation of a valid
>>> signature for a pre-defined public key and message or hash. The ledger can
>>> then easily validate the signature when it is presented and determine if
>>> the condition has been met.
>>>
>>> http://interledger.org/interledger.pdf
>>>
>>> Question: are signatures a necessary component in the ledgers?
>>>
>>
>


-- 
Evan Schwartz | Software Architect | Ripple
[image: ripple.com] <http://ripple.com>
Received on Tuesday, 26 January 2016 20:09:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 26 January 2016 20:09:48 UTC