Re: [Sat] stage0 v2 - message exchange diagram / SATP pre-conditions

Right. We need explicit return messages most likely via a return mechanism as call-back hooks (as messaging would be asynchronous). I will review and update 
Thank you for the note!
--
Denis

> On 21 Mar 2023, at 20:11, Thomas Hardjono <hardjono@mit.edu> wrote:
> 
> 
> Hi David and Denis,
> 
>>>> The problem with step 4 is that there appears to be no feedback to
>>>> client1App from client2App after receipt of the message.
>>>> If step 4 is not received by client2App steps 6,7, and 8 will not be
>>>> executed.
> 
> I have been looking at this diagram as a kind of a "sketch" (i.e. without the full request-response flow between the Applications).
> 
> So, Step-4 would be implemented with message transport reliability (i.e. it may be a 2 message sub flow).
> 
> 
> 
> --thomas
> 
> 
> ________________________________________
> From: Denis Avrilionis [denis@compell.io]
> Sent: Tuesday, March 21, 2023 1:09 PM
> To: ladler2@bellatlantic.net; Thomas Hardjono; Venkatraman Ramakrishna
> Cc: sat@ietf.org
> Subject: stage0 v2 - message exchange diagram / SATP pre-conditions
> 
> Attached is a revision of the Stage 0 diagram where I aligned the flow to the swimlanes of the core SAT message flow (thank you for your input!).
> 
> 
> Answering to (Claire’s I think) question on how we can translate this to a set of pre-conditions to initiate SATP I think we can state them as follows:
> 
> - The Originator Application (a.k.a. Client App A1, a.k.a. Alice) has obtained a “transfer context”, i.e. information that unambiguously defines a specific asset transfer instance related to sending a well-defined set of assets from a given Originator to a given Beneficiary
> 
> - Network 1 (a.k.a. State Data DB1) is aware that this well-defined set of managed assets (owned by the originator) is going to be transferred for the given transfer context, and this at any moment in the future, based on Gateway 1 own initiative.
> 
> - Gateway 1 is bound to Gateway 2 for the given transfer context
> 
> - The Beneficiary Application (a.k.a. Client App A2, a.k.a. Bob) knows that it will receive a well-defined set of assets for the given transfer context
> 
> - Network 2 (a.k.a. State Data DB2) is aware that assets will be received for a given beneficiary for the given transfer context
> 
> - Gateway is bound to Gateway 1 for the given transfer context
> 
> 
> Only if all the above conditions are met then the SATP transfer instance can be completed (note: completion can be either commit or rollback).
> 
> 
> 
> 
> On 21 Mar 2023, at 18:16, <ladler2@bellatlantic.net<mailto:ladler2@bellatlantic.net>> <ladler2@bellatlantic.net<mailto:ladler2@bellatlantic.net>> wrote:
> 
> Hi:
>  Yes it is progress that the stage0 message exchange defines a startup and
> a conclusion to the SATP process.
> 
> The problem with step 4 is that there appears to be no feedback to
> client1App from client2App after receipt of the message.
> If step 4 is not received by client2App steps 6,7, and 8 will not be
> executed. This would block the entire SATP process.
> Also when would client1App initiate step 5?
> 
> Also, if client1App and client2App are running on computers internal to a
> corporate network I don't see how network
> security  would allow anything except an email to go from the client1App
> computer to the client2App computer.
> 
> David Millman
> 
> -----Original Message-----
> From: sat <sat-bounces@ietf.org<mailto:sat-bounces@ietf.org>> On Behalf Of Thomas Hardjono
> Sent: Tuesday, March 21, 2023 10:38 AM
> To: Venkatraman Ramakrishna <vramakr2@in.ibm.com<mailto:vramakr2@in.ibm.com>>; ladler2@bellatlantic.net<mailto:ladler2@bellatlantic.net>;
> sat@ietf.org<mailto:sat@ietf.org>
> Subject: Re: [Sat] stage0 message exchange diagram
> 
> 
> Hi David,
> 
> What's interesting to me in Denis' applications-interaction diagram is that
> the gateway selection/election occurs (already completed) between Line-4 and
> Line-5 in Denis' diagram.
> 
> So by the time we start the SATP flow (the grey area) with a
> Transfer-Commence message 2.1, gateway G1 and G2 already "know" each other
> via Line-8.
> 
> (ps. I'm using the word "Line" for Denis's applications interaction diagram,
> and "Message" for our message flow)
> 
> 
> Although this applications-interaction is out of scope for SATP, it answers
> one of the questions we received last year (about showing a complete
> end-to-end flow).
> 
> 
> 
> --thomas
> 
> 
> 
> ________________________________________
> From: sat [sat-bounces@ietf.org] on behalf of Venkatraman Ramakrishna
> [vramakr2@in.ibm.com]
> Sent: Tuesday, March 21, 2023 6:59 AM
> To: ladler2@bellatlantic.net; sat@ietf.org
> Subject: Re: [Sat] stage0 message exchange diagram
> 
> My answers:
> 
> 1.  Yes, they refer to the same things. We should use consistent
> terminology here; I prefer either Ledger or State DB and not "system", which
> we use interchangeably with "network" in the charter and in the drafts.
> 2.  Step 4 is supposed to be oblivious to the gateway. I don't think we
> need to prescribe a standard transport mechanism, since it is out of scope
> for the current SATP, but I imagine we can use HTTP or GRPC.
> 
> Rama
> 
> From: sat <sat-bounces@ietf.org> On Behalf Of ladler2@bellatlantic.net
> Sent: 18 March 2023 23:56
> To: sat@ietf.org
> Subject: [EXTERNAL] [Sat] stage0 message exchange diagram
> 
> Hi: I have some questions on the Stage 0 message exchange: Are system1 and
> system2 the same as State Data DB1 and DB2 given in the SATP Message Flow
> Diagram? How is the step 4 message "propagateTransferContext" sent from
> client1App ZjQcmQRYFpfptBannerStart This Message Is From an Untrusted Sender
> 
> You have not previously corresponded with this sender.
> 
> 
> 
> ZjQcmQRYFpfptBannerEnd
> Hi:
>  I have some questions on the Stage 0 message exchange:
> 
> 
> 1.  Are system1 and system2 the same as State  Data DB1 and DB2 given in
> the SATP Message Flow Diagram?
> 2.  How is the step 4 message "propagateTransferContext" sent from
> client1App to client2App?
> 
> Does the message pass through gateway1 and gateway2?  If not what is   the
> underlying transport mechanism?
> 
> David Millman
> 
> --
> sat mailing list
> sat@ietf.org
> https://www.ietf.org/mailman/listinfo/sat
> 
> --
> sat mailing list
> sat@ietf.org
> https://www.ietf.org/mailman/listinfo/sat
> 
>