It is called the principle of encapsulation.
The most successful Internet protocols do not involve connections to hosts
today. SMTP is a connection to a service and has been for two decades. HTTP is
not quite so agile but would be had we had SRV at the time.
In SMTP the IP address does not remain constant end to end and never did.
Simply asserting that "there will still be some need to talk to a host or an
interface" without giving instances is hardly a compelling argument. More proof
by unsupported assertion seems to me.
From: Keith Moore [mailto:moore(_at_)network-heretics(_dot_)com]
Sent: Thu 11/13/2008 5:28 PM
To: Hallam-Baker, Phillip
Cc: Mark Townsley; Eric Klein; Routing Research Group Mailing List; Behave WG;
Subject: Re: [BEHAVE] Can we have on NAT66 discussion?
Hallam-Baker, Phillip wrote:
I beleive that the question would not arise If we had a coherent
The idea that an application can or should care that the IP address of a
packet is constant from source to destination is plain bonkers.
On the contrary, the idea that an application must not care that the IP
address of a packet is consistent from source to destination is plain
bonkers. Even assuming the existence of a higher level identifier and a
secure, fast, scalable, reliable way of finding routes to that
identifier, there will still be some need to talk to a host or an interface.
And nobody has demonstrated an application-independent mapping service
that is anywhere nearly up to that task. Until somebody does,
statements about what the architecture "should" do, or what applications
"should not" do with addresses, are at best wishful thinking, and at
worst delusion. And the more often someone makes such a statement
without qualification, the more it looks like the latter.
Ietf mailing list