[ih] Fwd: History from 1960s to 2025 (ARPANET to TCP)

[Brian E Carpenter]

Hi Tony, et al,

>  So far, no problem has crossed the pain threshold to make that effort worthwhile.

My understanding is that the pain that produces change is usually OPEX. It's certainly true that the IETF is mainly deaf and blind to OPEX issues. And also, a large part of the market doesn't perceive that their OPEX problems are caused by legacy protocol issues (CGNAT OPEX is certainly a case in point).

It will be interesting to see whether the DetNet (deterministic networking) stuff flies. That means explicitly moving connection management from the end host into the newtork.

In terms of IPv6 bells and whistles, the future is unsettled. Quite a debate is going on about draft-herbert-deprecate-eh which would deprecate most extension headers, and draft-herbert-deprecate-auth-header which would exterminate the IPSec authentication header. And IPv6 fragmentation is useless also.

I'd like to have a preview of what this list will be talking about in 2046.

Regards/Ngā mihi
    Brian Carpenter

On 04-Jan-26 13:41, Jack Haverty via Internet-history wrote:
> Forwarding for Tony Li -- email to the list doesn't work for him. /Jack
> 
> 
> -------- Forwarded Message --------
> Subject: 	Re: [ih] History from 1960s to 2025 (ARPANET to TCP)
> Date: 	Sat, 3 Jan 2026 12:58:06 -0800
> From: 	Tony Li <tony.li at tony.li>
> To: 	Jack Haverty <jack at 3kitty.org>
> CC: 	Brian E Carpenter via Internet-history
> <internet-history at elists.isoc.org>
> 
> 
> 
> Hi Jack,
> 
> 
>> In the new Internet architecture today, TCPV4 is still in use, even
>> though TCPV6 was "on the shelf" decades ago. Has the delay been a
>> result of the change in architecture? Are we missing the "process" for
>> evolution of the networking mechanisms? Is such a process even
>> possible given the size and breadth of the Internet?
> 
> 
> The delay comes from a very simple reason: a lack of a compelling reason
> to change.
> IPv6 is architecturally identical to IPv4, intentionally. However, this
> also means that IPv6 has no significant advantages over IPv4. Those that
> have been invented have been retrofitted to IPv4. We necessarily
> grandfathered all of the IPv4 Internet indefinitely into the future, so
> there is no compulsion to change and no desire to do so. Conversion is
> laborious process, with no significant upside.
> 
> 
>> So, my basic question for History is "Why did the architecture
>> change?" Were the arguments for a separate network switch (e.g., an
>> IMP) no longer applicable? Did the technology explosion during the 70s
>> have some effect? What was the reasoning behind the decision to move
>> the "virtual circuit" mechanisms from the network (IMPs) to the hosts?
> 
> 
> The architecture changed because it needed to. But because we are
> unwilling to consider architectural change now, we are stuck with where
> we have been since 1983. Moving connection management to the end host
> was an absolutely necessity for scalability and stability. That was the
> right call.
> 
> Moving forward, if we want to make architectural change and get it
> deployed, we have to find a way to motivate it so that it gets through
> the IETF process, through vendor implementation, and through end-user
> deployment. That implies that it has to solve a serious, real problem
> with the current architecture. So far, no problem has crossed the pain
> threshold to make that effort worthwhile.
> 
> Regards,
> T
> 
> p.s. For some reason, my postings never seem to make it onto the list.
> Please feel free to forward if you like.
> 
>