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

[Phillip Hallam-Baker <[e-mail]>]

On Thu, Feb 27, 2020 at 7:45 PM Keith Moore <[e-mail]>
wrote:

> On 2/27/20 7:21 PM, Phillip Hallam-Baker wrote:
>
> As the application layer designer, I am the customer here. I do not care
> about the IP address.
>
> You are only one of many, many application level designers.   The fact
> that you don't care about the IP address doesn't mean that no application
> designer needs to care about the IP address.
>
> Fundamentally the Internet is a peer-to-peer network, and there's no
> particular reason to assume that peers only interact in pairs.
>
> Keith
>
I am not aware of any successful application designers who design stuff
that doesn't work in the real world.

You have been on this jihad against NAT for decades. In the real world,
application designers have to accept NAT is simply a fact of life or their
stuff doesn't work. That IPSEC is a failure as a VPN standard is not an
opinion, it is a fact. Every VPN vendor developed their own work-around for
the AH debacle and as a result, the built in clients in Windows and Mac are
rarely able to connect to an IPSEC VPN. Meanwhile, SSH just works.

The job is to get the job done.

The failure of the Internet architecture is that there isn't actually a
document that describes it that is remotely close to being current. We
should have a document that describes the various network interfaces, the
interaction with the naming systems (BGP. DNS, URI) and what the upper
(i.e. more abstract) representation can rely on from the lower but we don't.