Re: [arch-d] ETSI launches new group on Non-IP Networking addressing 5G new services

On Thu, Apr 09, 2020 at 05:48:37AM -0700, Marie-Jose Montpetit wrote:
> Interesting that you mention vertically integrated solutions because in
> COIN and in related IoT research we are looking at common data layers to
> avoid having to go back to the DC or cloud before applications (in
> industry, agriculture etc.) can interoperate or data can be processed and
> communicated in order to create more advanced solutions using distributed
> cognition and machine learning.

Yes. interesting virtual meeting earlier this week.

> I know called these ???pet projects??? in a
> previous email but believe me there is not pet there but a real need.

Maybe i have not been using the term "pet" correctly, i was
meaning to have it express something people are passionate about.

Maybe i should have rather pointed to the financial incentives:
Folks contributing work to IETF and IRTF do this because its
most often part of their job/funding and its not really primarily
self-elected (beyond the fact how they may be able to self-elect
within bounds their funding). I think situation of topic selection by IAB
members follows quite different rules. For better or worse.

> And I gather you may know about the ICNRG since your team was contributing
> :)

Only had limied time to track that. Unfortunately a lot of
defunding of ICN going on i think.

> And while this thread has gone on for a while and
> it???s nice to criticize what are you exactly proposing.
> What is the architecture you would like to be discussed?

Well, this threas was not started by what architecure
i would like to see, that was a different thread, but
thanks for asking:

If you want to provide more explicit criticism for what wee
have put out so far, please comment on:

https://datatracker.ietf.org/doc/draft-bryant-arch-fwd-layer-ps/

Btw: while it will be fun to focus on the pieces you don't
like (common feedback has been "i do not believe in hologram
use-cases"), its obviously also helpful to comment on what
you do agree on.

This is not really proposals for specific architecture or protocol
components, because there are certainly a lot of possible options.

We started to write down the areas of concerns as a gap
analysis because we have primarily seen the discuss about
anything "New IP" come back with "what problem are you trying
to solve". Of course we have ideas and proposals how to solve
them, and will write them down too, but it would be a lot
more interesting if other members of the community would bring
in their ideas as well.

Practically speaking, in our opinions, most of interesting
future internet traffic that needs more than best-effort-network
services will happen in metropolitan networks between
towers/usrs and edge-DC, and i think we can and need to
evolve the internet architecture for that future on many
laers, starting at making the actual forwarding plane
of all routes/switches programmable (which would be where
i think your interest might be), up to the point of how
to introduce monetization models for consumption of such
future network services beyond the current model of
"flat rate for network transport and secret anticompetitive backend
deals for differentiated service delivery"

Cheers
    Toerless

> mjm
> 
> Marie-José Montpetit, Ph.D.
> marie@mjmontpetit.com
> 
> 
> 
> On April 9, 2020 at 8:24:33 AM, Toerless Eckert (tte@cs.fau.de) wrote:
> 
> To me, the interesting question is really if/whether ICN
> has a chance to happen as software in data centers.
> Last i looked (2 years ago ?), Cisco seems to have made
> that pitch architecturally.
> 
> I am always sceptical of classical interoperating
> networking solutions (ICN or else) to be successul
> in that DC software-only deployment model. Lot more
> abilities for vertical proprietary solutions to be
> faster successful.
> 
> Any good "state of affairs" summary in ICNRG ?
> Sorry... didn't track the RG.
> 
> Cheers
> Toerless
> 
> On Thu, Apr 09, 2020 at 12:07:26PM +0300, Lars Eggert wrote:
> > On 2020-4-8, at 21:57, Carsten Bormann <cabo@tzi.org> wrote:
> > > There has been a decade of ???clean slate Internet??? discussion, and
> only ICN emerged as an interesting new architecture from that context. So
> I???m not very optimistic with any process that tries to relive that
> process. But genuine new ideas do get generated now and then, and I would
> be quite interested to hear about them without the now obligatory press
> release bombast on how they will create cheaper battery technology for
> flying cars and cure Covid-19 at the same time.
> >
> > +1
> >
> > An even ICN has an uncertain road to larger-scale deployment.
> >
> > Lars
> 
> _______________________________________________
> Architecture-discuss mailing list
> Architecture-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/architecture-discuss