[ih] Packet nets not connected to the ARPAnet

[Louis Mamakos]

In the R&E space, there were certainly a number of non-IP based "internets" running around.  I recall NASA's SPAN, DoE ESNET and another small NSF-sponsered, "supercomputer" satellite-based network.  These were all based on DECNET, and it was an interesting challenge to plumb them into the network at the University of Maryland at the time, given the relatively small-size numbering space and lack of global uniqueness in how all the different members independently chose their network numbers.  I think there was also an MFENET out there at the time, too, also DECNET based.

Even with the early R&E IP networks, it was still a challenge to lash all these together in the early days given the non-existent to limited tools available for inter-domain routing that was available at the time.  Early attempts at using RIP were… not completely successful when the diameter of the interconnected routers exceeded "infinity" - defined to be 16.  EGP wasn't really suited for generalized network topologies, and it wasn't until BGP really came along that we had real solutions to that problem. 

At U of MD (like other sites at the time), we had interesting plumbing problems with connections to both the MILNET and ARPANET (in different departments on campus, just to make it interesting) as well as being an early NSFNET site, and a NSF regional network (SURANET) site; as well as internal university system networks to multiple campus sites.  Many lessons were learned, horrific kludges perpetrated.  gated experienced quite a bit of violence and insult, but we felt our way through what was needed/possible to make interdomain routing work, in a way that was consistent with the funding/cost models related to the various attached networks.

And yeah, Ethernet is a protocol.  I recall reading the Pascal pseudo-code in my DEC-Intel-Xerox Ethernet specification.

Louis Mamakos


On Jul 23, 2012, at 7:18 PM, Alex McKenzie wrote:

> Yes, There were packet networks not connected to the ARPAnet.  Two in particular were: Cyclades/Cigale in France, and the European Informatics Network with nodes in 6 of the Common Market countries.  There _were_ some interconnections between the public packet networks (Telenet in the USA, Bell Canada, and the networks of the PTTs in western Europe and Japan), but these interconnections were extremely limited experiments for the most part.  CSNET (funded by the NSF) used both ARPAnet and Telenet for transport but only provided user-level (eg mail) interconnection and only for CSNET members.
> 
> From: Ofer Inbar <cos at aaaaa.org>
> To: Ian Peter <ian.peter at ianpeter.com> 
> Cc: internet-history at postel.org 
> Sent: Monday, July 23, 2012 6:41 PM
> Subject: Re: [ih] FYI - Gordon Crovitz/WSJ on "Who Really Invented the Internet?"
> 
> Ian Peter <ian.peter at ianpeter.com> wrote:
> > The writer of the LA Times article might have done well if he had checked
> > with his friend Bob Taylor. He writes.
> > 
> > "Bob Taylor is a friend of mine, and I think I can say without fear of
> > contradiction that he fully endorses the idea as a point of personal pride
> > that the government-funded ARPANet was very much the precursor of the
> > Internet as we know it today"
> > 
> > Well that's not what Bob Taylor has written elsewhere. To quote Bob,
> > 
> > "I believe the first internet was created at Xerox PARC, circa '75, when we
> > connected, via PUP, the Ethernet with the ARPAnet. PUP (PARC Universal
> > Protocol) was instrumental later in defining TCP."
> 
> That supports the assertion that the ARPANet was the precursor of the
> Internet, so I see no contradiction.
> 
> Note that "the first internet" per se wouldn't necessarily have to be
> a precursor of "The Internet".  Were there any packet switched
> internets* in those days that did not connect to the ARPANet?  While
> those would not have had a direct lineage to The Internet, it'd still
> be interesting to know about them.  I don't recall hearing of any.
> 
> * by which I mean: interconnections of administratively separate
>   networks in geographically separate locations with different
>   underlying network types - administratively separate is key
> 
>   -- Cos
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20120723/55b8b4a0/attachment.htm>