Re: [nbs] Preparing for a BoF about name-based sockets.

[Dave Thaler <[e-mail]>]

> On 2010-09-15 07:17, Javier Ubillos wrote:
> > Hi everyone.
> >
> > We think we got some good feedback after the two presentation at
> > Maastricht, so we've decided to move towards a BoF.
> >
> > We are writing to you to ask for some input and feedback.
> >
> > We would like to talk some about ID/Loc (names & locators) in general
> > and explore if there is more room within the IETF for it.
> > The short version is: Today we have ID/Loc splits at different levels,
> > however we think that are some advantages to this one.
> >
> > Relevant documents:
> > BoF Proposal: https://www.sics.se/nbs/bof/Proposal%20for%20Name-
> Based%
> > 20Sockets%20BoF.pdf
> > Draft: http://www.ietf.org/id/draft-ubillos-name-based-sockets-02.txt
> >
> > About the draft: Please note, the draft _will_ be divided into two
> > different documents, one about name based sockets, and one about how
> > shim6 can be improved with name-based sockets.
> > We have been a bit unclear earlier, name-based sockets is about the
> > naming layer & mechanisms. The described shim6 improvements are an
> > example on solutions that can be improved and integrated using
> > name-based sockets.
> >
> > // Christan Vogt and Javier Ubillos

I support having a BOF on this topic.

I think splitting the draft is a good thing, and I have feedback on at least the
API portion.  Specifically, the API needs to use service names, not integer
port numbers.   getaddrinfo() already does so, and section 7.2 of
draft-ietf-tsvwg-iana-ports-06 will only make this more important.

Don't make the mistake of using integer port numbers in an API that
claims to be "name-based".

-Dave Thaler