[ih] History of Tier 1 Networks

[Greg Skinner]

On Apr 26, 2026, at 5:02 PM, Tony Li via Internet-history <internet-history at elists.isoc.org> wrote:
> 
> 
> Hi Brian,
> 
> [As always, my answers never make it to the list. Apparently I am on the wrong side of a diode.  Please feel free to forward if it moves you.]
> 
> 
>> On Apr 26, 2026, at 4:43 PM, Brian E Carpenter via Internet-history <internet-history at elists.isoc.org> wrote:
>> 
>> James,
>> 
>>> The main question I want to understand is how the current set of Tier 1 networks came to be Tier 1 networks (why these networks and not a different set of networks?).
>> 
>> I suspect that the answer will turn out to be the same as for any other industry since the Industrial Revolution: some combination of effective entrepreneurship and pure chance. Did a particular operator (a) understand the strategic value of being a transit provider for no *direct* financial reward and (b) have good enough technical staff and (c) have a bit of luck? Not to mention freedom from political interference.
> 
> 
> The primary driver for this is, as always, economics. You’ll note that these Tier 1 operators are all former PTTs and telcos.  They were positioned well because they had already invested in the primary resource: installed fiber.  When the ISPs blossomed, it became immediately apparent that the networks that had installed fiber (a.k.a., facilities based) had a huge economic advantage: they got bandwidth for the cost of right-of-way, installation, and maintenance.  ISPs that were not facilities based had to go find bandwidth on the open market, and prices reflect not only the costs above, but a hefty markup thanks to the supply/demand skew.
> 
> The other driver, of course, was money.  Telco’s decided to get into the game and immediately started acquiring regional networks, thanks to their deep pockets.
> 
> 
>> But don't neglect the role of IXPs in this. Without them, nothing would work. How and why did they come into existence?
> 
> 
> Again economics. Just as carrier hotels were common in the pre-Internet telco era, ISPs found it more economical to drop a line and a router to an IX than to engineer individual circuits to each of the networks that they wanted to peer with.  As traffic volumes have risen, some of the economics have shifted away from this, and so suitably large peerings have shifted to direct private fiber.  FIX-East and FIX-West were the initial instantiations, which in turn evolved into MAE-East and MAE-West and now to commerical IXPs.  For more on the economics of peering and IXPs and why the IXPs need to be provider independent, I refer you to William Norton’s “The Internet Peering Playbook”.
> 
> Regards,
> Tony
> 

I’m copying WBN on my reply as IMO, he’s one of the most knowledgeable people on this and related subjects.  He has a site, Ask Dr. Peering <https://drpeering.net/HTML_IPP/index.html>, where information like this is available, including a link to the Playbook.

Some other possible sources of information (which may require deep dives through old email and Usenet posts):

https://web.archive.org/web/19980118011753/http://www.merit.edu/mail.archives/mjts/ (Merit Joint Technical Staff email archives from 1991-1997)
https://groups.google.com/g/mlist.com-priv (Usenet gateway of the com-priv mailing list from 1995-2000, incomplete)
http://www.telecom-digest.org/ (Telecom Digest, starting in 1981)

--gregbo