Re: [icnrg] reference needed for Interest-Interest

Marc,

Thanks!

Dirk

On 9 Mar 2020, at 18:23, Mosko, Marc wrote:

> The interest-interest-data semantics have been around in CCNx since 
> the very early days.  It is the CCNx 0.x repository StartWrite command 
> (https://github.com/ProjectCCNx/ccnx/blob/master/doc/technical/RepoProtocol.txt). 
>  A producer sends an Interest to the repo asking that it fetch data 
> from the producer.  The Repo sends an acknowledgement ContentObject 
> and also starts fetching the data via normal interest-data exchanges.
>
> The main issue with this approach is (a) the producer cannot tell 
> which interests are from the repo and which are from another consumer, 
> and (b) there is no "finished" notification, so the client must poll 
> the repo to see if it has everything.  Both (a) and (b) mean the 
> producer cannot easily discard the content as it does not know when 
> the repo is done saving it.  A client can check if a particular 
> segment is in the repository via the CheckedStartWrite command, though 
> this is done segment-by-segment so is not particularly efficient.
>
> The early Voice over CCNx (2009) paper used an Interest-Data exchange 
> to kickoff bi-directional Interest-Data/Interest-Data exchanges in a 
> different namespace.
>
> The paper "CCNx 1.0 Bidirectional Streams " 
> [https://arxiv.org/abs/1707.04738] extends the VoCCNx approach to a 
> more robust repository protocol.   The document was originally 
> published in 2013 or so as a tech report with the early CCNx 1.0 
> papers.
>
> Marc
>
> On 3/9/20, 7:12 AM, "icnrg on behalf of Dirk Kutscher" 
> <icnrg-bounces@irtf.org on behalf of ietf@dkutscher.net> wrote:
>
>     https://conferences.sigcomm.org/sigcomm/2011/papers/icn/p56.pdf
>
>     from ICN-2011 describes the approach and its potential problems.
>
>     Let me know if you have a better (seminal) reference.
>
>     Thanks,
>     Dirk
>
>     On 9 Mar 2020, at 14:21, Dirk Kutscher wrote:
>
>     > Hello everyone,
>     >
>     > I’m looking for a reference for the 
> “Interest-Interest-Data”
>     > scheme (that has been proposed as a PUSH surrogate).
>     >
>     > Thanks,
>     > Dirk
>     >
>     > _______________________________________________
>     > icnrg mailing list
>     > icnrg@irtf.org
>     > https://www.irtf.org/mailman/listinfo/icnrg
>
>     _______________________________________________
>     icnrg mailing list
>     icnrg@irtf.org
>     https://www.irtf.org/mailman/listinfo/icnrg