[ih] X.25

[Jack Haverty]

It's also important to remember that X.25 was only an "interface 
specification" for how a computer could interact with a network. Inside 
the network, mechanisms were defined by the PSN designer. As far as I 
remember, an X.25 network always consisted of switches from the same 
manufacturer or possibly partners.   X.75 was used to connect two 
networks together.

For example, the US DDN used X.25 as the specification for how a host 
computer would use the DDN.   But inside the DDN switches (aka IMPs) 
were the same mechanisms (routing, flow control, congestion control, 
management, etc.) for creating reliable virtual circuits that had been 
developed and refined over a decade of use in the ARPANET.  Only the 
host/switch interface had changed.

In the 1980s, here were multiple distinct X.25 networks as components of 
DDN, as well as similar clones in other parts of the US government.  I 
don't recall that they were ever interconnected by X.75 or even by IP 
gateways, but it may have happened somewhere.

There were some "mail gateways" that enabled email to move between some 
networks, using trusted agents interfaced to multiple networks.  An 
example is documented in the RFCs - see 
https://datatracker.ietf.org/doc/html/rfc1133

One memorable example I recall was a "mail gateway" in Europe between 
the DDN and the European public X.25 system - in Germany IIRC.   Because 
of security concerns, that mail gateway was implemented organically -- a 
human operator at a desk with two terminals would read a message on one 
and, if it was acceptable, re-type that message on the other terminal 
and send it to its destination.

This connectivity was useful for mundane needs like getting groceries, 
toilet paper, and other consumables ordered and delivered to military 
installations.   We used to joke that an alternative implementation 
could be a gateway between the DDN and some Avian Network but no ISP 
offered such service, and it was easier to find a soldier to be the gateway.

/Jack

On 10/2/25 13:00, Brian E Carpenter via Internet-history wrote:
> On 03-Oct-25 06:41, Lyndon Nerenberg (VE7TFX/VE6BBM) via 
> Internet-history wrote:
>> John Levine via Internet-history writes:
>>
>>> I was a long time uucp user and I do not ever recall running
>>> into anyone who used it over X.25.
>>>
>>> This web page says that the widely used Taylor uucp
>>> had an 'x' protocol for X.25 but also says it doesn't work:
>>
>> Berkeley (I think?) added the 'f' protocol for use over X.25 via
>> PADs.  It encoded 8-bit traffic into 7-bit values, used xon/xoff
>> flow control, and probably had a few other quirks I forget.  And
>> it was certainly used, at least in Canada.  Before CA*net was formed,
>> the U of Alberta, UBC, U of Waterloo, and probably U of Toronto,
>> all swapped mail and Usenet over Datapac via UUCP running 'f'. When
>> I set up APSS (Alberta Disaster Services), I set up a UUCP link
>> over Datapac with the UofA running 'f' protocol to relay news amd
>> mail.  It was a real thing.
>
> The irony being that Datapac was merely an X.25 wrapper on an underlying
> connectionless packet-switched network [1] [2].
>
> As others have said, IP over X.25 was fairly common in Europe, for 
> financial
> or political reasons. In fact several European efforts at 
> everything-over-X.25
> can be found in the two relevant history books [3] [4].
>
>    Brian
>
> [1] 
> https://www.cs.auckland.ac.nz/research/groups/CDMTCS/researchreports/download.php?selected-id=884
> [2] https://ieeexplore.ieee.org/document/6592834
> [3] 
> http://ictconsulting.ch/reports/european-research-internet-history.pdf
> [4] https://doi.org/10.1002/9783527629336

-------------- 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/20251002/2a08f166/attachment.asc>