Re: [arch-d] IAB Technical Discussion on Fragmentation: 2023-05-03

[Watson Ladd <[e-mail]>]

On Tue, May 2, 2023, 3:38 AM Vittorio Bertola
<vittorio.bertola=[e-mail]> wrote:
>
>
>
> Il 28/04/2023 23:54 CEST Andrew G. Malis <[e-mail]> ha scritto:
>
>
> Bob,
>
> I also had the same initial thought. And I've even got an RFC discussing pseudowire fragmentation, so I was looking forward to an IP fragmentation discussion!
>
> I guess this is part of the current, extensive effort by the I* organisations (mostly ISOC and ICANN) to counter any renewed proposals for a different governance structure for the key technical resources of the Internet, i.e. names and numbers (did anyone say ITU?).
>
> However, there is some ambiguity on whether this also extends to regulation at the application layer, like the recent efforts in Europe to establish rules that would force the global Internet industry to tailor their services to specific European norms (this may go as far, to give you an example of a current discussion, to require AI models to be able to forget all information they learned about a specific European citizen, which of course is not technically easy; but it also includes pro-competition measures against walled gardens, requirements for large online information platforms, and, more controversially, proposals to mandate the screening of personal messages for CSAM).
>
> I am curious to see how the IAB defines "fragmentation" - this is an ongoing discussion at the already mentioned IGF PNIF - and thus what is the scope of this "technical discussion". Personally, I would encourage the IAB to stay out of the latter set of problems, as there is almost nothing technical in them - it's all about policy, values, interests and compromises.

I seem to remember two WGs facing hostility due to policy and values
issues: ohai and privacypass, and then the W3C has had its own issues
with the Privacy Sandbox. So I don't think we can ignore these
supratechnical issues entirely if participants are saying we should or
shouldn't charter based on them. Also, ODoH, etc.

This isn't new: we saw these sorts of issues in the 1990's with
encryption, the 2010's post Snowden where the choices of what to
standardize had a political valence.  Putting our head in the sand
when participants are discussing these issues and bringing them up as
reasons to standardize or not strikes me as counterproductive. It's
also worth remembering that there is no protocol police: you can
deploy any application to do anything on the net, and provided it
doesn't cause problems for the rest of the network, or really problems
for anyone who sits between you and the other users, it works. As a
result standardization isn't really able to do anything more than say
"no, that's a bad idea", but isn't able to actually stop it.

The fact that the Internet is permissionless and implementation
precedes standardization is good: it creates a dynamic environment
that has tremendously benefited everyone involved, and has prevented
rent-seekers who would like to stop investing and innovating and
instead charge tolls.  That wasn't just a policy choice, but a result
of technical decisions that enabled a decentralized network with
distributed namespaces, as well as norms around interconnection, and I
think that's more than just ISOC. And it also means that what the
future of the Internet is is not controlled by the IETF but rather
shaped by commercial forces and responses to emerging
needs+technologies.

Sincerely,
Watson

>
> Ciao,
>
> --
>
> Vittorio Bertola | Head of Policy & Innovation, Open-Xchange
> [e-mail]
> Office @ Via Treviso 12, 10144 Torino, Italy
>
> _______________________________________________
> Architecture-discuss mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/architecture-discuss