[ih] Failures of the early Internet

[Jack Haverty]

Hi Scott,

You might be right.   I still have a vague recollection that the two 
sites were somewhere in the SouthEast US, perhaps Carolinas or Georgia.  
But I'm sure that the incident I remember was not unique. One thing we 
learned about routing was that slight changes could have dramatic and 
unexpected effects on how the overall system behaved.

At the time of your story, EGP was probably in effect and the "core" 
gateways insulated from "fake news" from the world outside.  So the 
"service" part of the Internet was protected from, yet still 
interoperable with, the "research" parts of the Internet, with all 
pieces of The Internet insulated into a handful of separate Autonomous 
Systems.

That was the intent and goal of the concept of Autonomous Systems and 
the EGP protocol - to keep the core service reliable while enabling lots 
of experimentation (especially in the hyperactive Fuzzy Universe), to 
resolve the outstanding issues and fold the resulting new algorithms and 
protocols into the next generation of TCP/UDP/IP/ICMP/GGP/etc.

Of course, that's not what happened.  ASes became a basis for growth 
instead, and there are now thousands of them.  The Internet followed its 
own path.

Jack Haverty

On 1/20/24 10:50, Scott Brim wrote:
> On Fri, Jan 19, 2024 at 9:20 PM Jack Haverty via Internet-history 
> <internet-history at elists.isoc.org> wrote:
>
>     Another incident I recall was also a routing issue.  I don't remember
>     exactly where it happened, but two sites, universities IIRC, were
>     collaborating on some research project and had a need to send data
>     back
>     and forth.  Their pathway to each other through the Internet was
>     somewhat long and often congested.   So they decided to fix the
>     problem
>     by installing a circuit directly between their two campus' routers.
>
>     Money was of course an issue, but they found the funds to pay for
>     a 9.6
>     kb/s line.  They were surprised to observe that the added line
>     only made
>     things worse.  File transfers took even longer than before. Of course
>     their change to the topology of the Internet had unexpectedly made
>     their
>     9.6 line the best route for all sorts of Internet traffic
>     unrelated to
>     their project.
>
>
> That might have been us, and if so it's another tie-in for Dave Mills. 
> In early January 1987 we lit up the first link in what was to be NYNET 
> (New York), between Cornell Theory Center and Columbia IT. We figured 
> Cornell would be a gateway for all of NYNET to the budding NSFNet. 
> However, at that time both Cornell and Columbia CS were connected to 
> ARPAnet, and Columbia CS was announcing a static route to HP (net 16) 
> to its campus, for some project in a department. At Cornell we 
> believed everything we received, so we forwarded Columbia's route to 
> the rest of our campus, thus to our CS Arpanet connection, and onward. 
> There was no route filtering anywhere. We discovered the HP routing 
> loop pretty quickly and shut down dynamic routing. Acouple weeks later 
> we were meeting with Dave, probably at UDel, and he said "we have to 
> have bidirectional route filtering", and thus the gated project was born.
>
> Scott

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 665 bytes
Desc: OpenPGP digital signature
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20240120/e96017d5/attachment.asc>