Re: [External] Request for information - Challenges in routing related to semantic addressing

[Toerless Eckert <[e-mail]>]

Thanks, Albert

You touch on more than one point.

I was primarily pointing to a technical opportunity that evolves the more transit
becomes required on paths. That does not mean that the opportunity is picked up
by players.

IMHO, 5G/6G and WiFi will evolve as additional competitive access options the
more they get built out. Yes, some countries such as the USA so far have not created 
regulation for a more competitive market but allowed operators to be quite uncompetitive.
That is also why customers in those markets are being extorted
on pricing (just based on limited personal experience, not a scientific analysis ;-)

Figuring out how to make use-case deployments benefitting from network services
beyond BE is a complicated issue. Vertical integrated services have in the past been
prime drivers. Slicing/virtualized-programmable network may be future enablers.

Cheers
    Toerless


On Wed, Mar 03, 2021 at 09:20:27PM +0000, Manfredi (US), Albert E wrote:
> -----Original Message-----
> From: routing-discussion <[e-mail]> On Behalf Of Toerless Eckert
> 
> > I would very much hope that there will still be competition in the Internet access,
> even if i also believe this will require more and more regulation given how the
> economics of scale work against a competitive market, but i think more and more
> countries are awakening to that challenge, for example through models of
> non-provider owned access fibers.
> >
> > So, when there are multiple access-provider and multiple edge-cloud providers,
> i hope there is enough inter-provider interop requirements to keep IETF relevant,
> but not enough to have the (IMHO) service-killing transit domain in the game.
> >
> > Aka: IMHO, the non-profileration of anything but best-effort service into "Internet"
> is really because of transit. If the future dominant paths e.g.: in metro areas
> would become:
> >
> > (subscriber/edge-cloudA) -- network-providerA -X- network -providerB -- (subscriber/edge-cloudB)
> 
> I don't know what the actual situation is in most countries. However, for essentially unmetered, cabled Internet access, in the US, it is mostly the case of only one "last mile" access ISP being available to a given household, two in many urban and suburban environments, and infrequently three. Which says to me, competition on the basis of "ATM-like class or quality of service" knobs, provided by your ISP, won't help much, unless everyone uses the same techniques?
> 
> Historically, what gave IP the edge over competing standards was, just throw cheaper bandwidth at the problem, available with IP's best effort model, and those tricky and expensive knobs become just an academic exercise. Are we saying that the situation has changed?
> 
> Bert

-- 
---
[e-mail]