[ih] Why did DCCP fail to gain traction?

[Scott O. Bradner]

at some level, DCCP was designed to solve the problem (middle boxes blocking non tcp/udp protocols - Quick (RFC 9000)
has done a fine job of solving that (and evolving the net to answer Jack’s last question

Scott

> On Oct 3, 2026, at 4:58 PM, Jack Haverty via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> In addition to technical issues, I think there was another reason for "failing to gain traction".  It comes down to management, or lack thereof, of new technology in an evolving infrastructure service.
> 
> Personally, I'm just one of billions of users of the Internet now.  But I was pretty heavily involved in the technology during the early years.   Today, my home LAN is an eclectic mix of devices, all on the Internet.  There are desktops and servers, running Windows, IOS, or Ubuntu.   There are phones, tablets, both Android and Apple systems, and even a few Raspberry Pis.   Several irrigation controllers and other home automation stuff.  Probably a bunch of things I've forgotten.  Even the fan in my attic is on the Internet.
> 
> I have no idea whether or not any of those devices use, or even are capable of using, DCCP.  DCCP is not alone - As just a user, I haven't heard about SCTP either, or many of the other technology that the IETF has documented in RFCs and placed "on the shelf" for anyone to use.  I've heard a little about QUIC, but don't know how to tell if any of my devices can or do use it.  Or why they should.  I suspect most of the billions of Users have similar experiences.
> 
> Looking outward, I see lots of ads for new devices to attach to my LAN.  But I can't recall ever seeing any ad that highlighted some product as "Supports DCCP!!".   Or "QUIC!", or "SCTP!", or even "IPv6!".   Companies traditionally tout the benefits of their products over the competition, and implement the features that their customers demand.  For some reason that doesn't seem to happen for Internet technology.
> 
> Back in 1983, a major effort was involved to transition the existing ARPANET from its old technology (called NCP) to the new TCP mechanisms.  That was accomplished on January 1, 1983, after several years of planning and preparing.  After that date, TCP was used everywhere, and NCP was no longer present anywhere.
> 
> Afterwards, we realized that such transitions were a lot of work, and that particular transition was greatly aided by the fact that the entire Internet at the time was owned and managed by a combination of ARPA and DCA, i.e., part of the US Department of Defense.   They were definitely "in charge."
> 
> We also realized that such a scheme wouldn't have the same advantage as the Internet grew and many more players got involved.  No one would be "in charge" and have the ability to orchestrate the introduction of new technology, as well as the removal of the older technology which it replaced.  Without both steps, the technical machinery would just get more and more complex and difficult to operate.
> 
> But no one in 1983 knew how to do that for the Internet.  It was possible that some lessons could be learned from the experience of the telephony system, which had pretty much succeeded at creating a multi-vendor and global voice service.   But the topic was still viewed as a research area, likely involving considerable research and trial and error to find some techniques that would work to not only introduce some new technology, but also to remove any old technology that was no longer appropriate.
> 
> The question for the research community was "How do you build a multi-vendor global communications system that can evolve with new protocols, algorithms, and techniques, and removal of the old technology that are to be replaced?"
> 
> As far as I can tell, that hasn't been answered yet.   It's been more than 40 years.  Maybe it's just not possible?
> 
> /Jack Haverty
> 
> 
> On 10/3/26 09:46, Thomas Fossati via Internet-history wrote:
>> On Sat, Oct 3, 2026 at 6:26 PM John Kristoff via Internet-history
>> <internet-history at elists.isoc.org> wrote:
>>> DCCP (IETF RFC 4330) is/was a relatively new transport in Internet
>>> history. I don't think it ever really saw much, if any, deployment or
>>> support in applications. It was removed from the Linux kernel last year.
>>> 
>>> Was it too late, too complex, too little, or just deemed unnecessary
>>> given available alternatives such as QUIC?  I don't recall seeing
>>> much discussion about why it failed to gain traction and I'm interested
>>> to hear what the collective wisdom of this group might have to say about
>>> that.
>> It failed because of an abundance of middleboxes and a lack of APIs.
>> Unlike SCTP (which is similarly affected by middleboxes), it never
>> found a niche where it could be safely deployed.
>> Then came QUIC, which made it completely redundant.
>> 
>> That's my rough understanding of the situation.
>> 
>> cheers!
>> 
>>> John
>>> --
>>> Internet-history mailing list
>>> Internet-history at elists.isoc.org
>>> https://elists.isoc.org/mailman/listinfo/internet-history
>>> -
>>> Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>> 
>> 
> 
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe: https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history