[ih] AUP revision to allow commercial traffic

[Brian E Carpenter]

On 23/12/2015 07:28, Jack Haverty wrote:

...
> Does anybody know more about the *kinds* of interconnections that were
> legal, illegal, feasible, and/or commonly used back in those early days?
>  E.G., was there any distinction between full IP connectivity (pass IP
> datagrams), versus limited protocol connectivity (perhaps FTP/SMTP
> only?). versus single-purpose connectivity (email only)?

I don't recall any intermediate steps between mail gatewaying and full
IP layer interconnection. There was a lot of talk about application
layer gateways other than email, but not much action. At CERN, everything
we did was deemed to be for research purposes, even if it involved commercial
partners, so we never worried too much about those pesky AUPs (except when
asking for funding, when we had to swear it was all for physics).

I guess the period when we had trouble sending email to uk.oracle.com
(which was sometimes assumed to be in JANET order and was automatically
flipped to com.oracle.uk), or to cs.ucl.ac.uk (sometimes flipped
to uk.ac.ucl.cs, i.e. Czechoslovakia) was just before full IP connectivity.

Another fragment is that AUP issues were discussed in the CCIRN (coordinating
committee for international research networking) from April 1989, if not
earlier. There were discrepancies between different countries and that
was a problem. Once we got fully into IP layer connectivity it all became
rather silly, but of course it evolved into BGP-4 policy in due course,
because nobody wanted to carry unfunded traffic.

    Brian