Re: [Int-area] Is IPv6 End-to-End? R.I.P. Architecture? (Fwd: Errata #5933 for RFC8200)

[Keith Moore <[e-mail]>]

On 2/27/20 5:07 PM, Tom Herbert wrote:

> I think we need to be careful that IETF is labeled as a collection of
> inflexible architectural purists.

There are several problems with the "we should not be architectural 
purists" argument:

1. It presupposes that there are no significant technical justifications 
for different components of the infrastructure, or different stacks or 
applications, to behave consistently with one another, when in fact that 
very consistency is what permits interoperation at all.

2. IETF's job is not to specify what makes some vendors happy, but to 
specify what is believed to work well and interoperate, not merely in 
the short term but also in the long term.   Vendors' interests and the 
interest of the Internet community are not the same thing.   While 
everyone understands that a specification that nobody implements serves 
no useful purpose, a specification that abandons good sense in order to 
please (some) vendors does harm to the Internet in the long term by 
adding more complexity, more operational issues, and more costs that are 
ultimately borne by users.

3. The perceptions of people outside the IETF should not be considered 
more important that the technical judgment of those who actually have to 
make the compromises that go in the specifications.   This is not to say 
that IETF always knows better than everybody else, but rather than the 
working groups and authors who actually write the specifications have 
some responsibility to make good compromises, whereas external critics 
bear no such responsibility.

4. There will always be people insisting that IETF consists of 
inflexible architectural purists, because this is a standard trope that 
can be used by anyone who objects to any consensus decision that IETF makes.

Keith