Re: [Spud] questions on the BoF outcome

[Spencer Dawkins at IETF <[e-mail]>]

Hi, Toerless,

On Fri, Jul 22, 2016 at 4:57 PM, Toerless Eckert <[e-mail]> wrote:

> Whats actually the official process for asking for support - only
> Hum i the WG, or is there going to be a call here or whatever other
> mailing list ?


Thanks for asking.

I met with the BOF proponents for at least a couple of hours today/tonight,
and talked with them about next steps.

I don't plan to ask for opinions about the charter as proposed at this
time, in order to give the proponents and chairs time to take those next
steps (they have two and a half hours of opinions from about 250 people to
absorb now).

If PLUS moves forward, the complete process would be something like

   - proponents absorb feedback to date
   - proponents take steps based on that feedback
   - proponents hand Spencer a revised charter, a more formal problem
   statement, and whatever else seems appropriate
   - Spencer asks the IESG and IAB for comments ("internal review")
   - Spencer absorbs comments received from internal review
   - Spencer takes steps based on those comments
   - The IESG asks the community for comments ("external review"). This
   includes both the IETF community and other SDOs

(This is where you come in :-)

   - The IESG absorbs feedback we receive from external review
   - Spencer takes steps based on that feedback
   - Spencer places the resulting charter on an IESG telechat agenda for
   approval of WG creation

I hope that's helpful! And thanks for asking.

Spencer


> If anyone is counting, here's one in favor of Plus WG.
>
> Use-case: Controlled network edge to Internet firewalls wanting to have
> similar
> flow recognition to TCP for UDP flows with replay-safe intent to receive
> indication.
>
> Cheers
>     Toerless
>
> On Fri, Jul 22, 2016 at 10:48:47AM -0400, Ian Swett wrote:
> > I hummed no largely because I felt I didn't clearly understand the work
> of
> > the group, and hence couldn't evaluate whether it was possible and
> > desirable.
> >
> > I personally would have preferred a clearly scoped set of initial use
> > cases, with other use cases of potential future interest requiring a
> > re-charter.
> >
> > I believe there are a lot of people interested in something along these
> > lines, so my no hum was not an expression they should stop trying to move
> > forward with a WG.  I also want to continue having the conversation about
> > what the path needs at the IETF, because it comes up fairly often, but I
> > feel there's currently no coherent forum for the conversation.
> >
> > It would be great to see some amount of running code as well, with some
> > clearly improved metrics for that particular use case.  I would hope
> small
> > scale experiments would inform the WG on what topics may warrant a
> > re-charter.
> >
> > On Fri, Jul 22, 2016 at 6:29 AM, Szilveszter Nadas <
> > [e-mail]> wrote:
> >
> > > Hi,
> > >
> > > Some other argument was e.g. overhead. There was a significant amount
> of
> > > "NO" hums for the first question, the rest of the questions was not
> even
> > > asked.
> > >
> > > It is also not clear to me, is there still hope, or is it hat eating
> time?
> > > :)
> > >
> > > Cheers,
> > > Sz.
> > >
> > > > -----Original Message-----
> > > > From: Spud [mailto:[e-mail]] On Behalf Of Fossati,
> Thomas
> > > > (Nokia - GB)
> > > > Sent: Friday, July 22, 2016 12:26
> > > > To: [e-mail]
> > > > Subject: [Spud] questions on the BoF outcome
> > > >
> > > > Hi,
> > > >
> > > > Unfortunately I could not attend the PLUS BoF.  So, I've just gone
> > > through the
> > > > minutes [1] (thanks a lot, scribes) and got the feeling that this
> work
> > > is pushed
> > > > back due to the perception that it'd weaken users' privacy?
> > > >
> > > > I hear these arguments:
> > > > - "potential to compel clients to send metadata or packets will
> dropped"
> > > >
> > > > But that could have happened already if the network wanted to (just
> drop
> > > any
> > > > TCP payload that starts with 0x16 and allow only clear-text
> traffic!).
> > > >  Access networks that you pay for do not have that incentive though,
> so
> > > I'm
> > > > very skeptical this could now happen *because of* PLUS.
> > > >
> > > > - "possibility for abuse"
> > > >
> > > > Well, that depends on the metadata that *users* decide to leak
> (which is
> > > a
> > > > separate discussion on the vocabulary), but in general Brian's
> framework
> > > looks
> > > > pretty well designed to bias control towards the endpoints which can
> act
> > > as
> > > > circuit-breakers at any point in time.
> > > >
> > > > - "giving more power to the network";
> > > >
> > > > This is actually true, but in a good way: the network will have
> power to
> > > send
> > > > useful information to the endpoints -- if it's asked to -- while
> being
> > > empowered
> > > > by the signalling coming from the endpoints (e.g., for DDoS
> prevention).
> > > >
> > > > So, sorry but this looks a lot like FUD to me.
> > > >
> > > > Is the working group not formed on these grounds?  Or have more
> > > substantial
> > > > weaknesses been highlighted during the discussion that have not been
> > > > captured in the minutes?
> > > >
> > > > Cheers, thanks,
> > > > t
> > > >
> > > > [1] http://etherpad.tools.ietf.org:9000/p/notes-ietf-96-plus
> > > >
> > > > _______________________________________________
> > > > Spud mailing list
> > > > [e-mail]
> > > > https://www.ietf.org/mailman/listinfo/spud
> > >
> > > _______________________________________________
> > > Spud mailing list
> > > [e-mail]
> > > https://www.ietf.org/mailman/listinfo/spud
> > >
>
> > _______________________________________________
> > Spud mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/spud
>
>
> --
> ---
> Toerless Eckert, [e-mail]
>
> _______________________________________________
> Spud mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/spud
>