[ih] AUP revision to allow commercial traffic

[Guy Almes]

One other perspective: in the late 1980s I directed one of the 
NSFnet-related regional networks -- this one in Texas.  Initially, all 
our members were universities and research centers.  Then we connected 
the research department of one of the major oil companies.  The 
motivation was clear, at least at first: to enable collaboration between 
industry researchers and university folks in departments such as geology 
and petroleum engineering.  We knew that if we thought about this, 
closed our eyes, and said "transitive closure" three times, something 
complicated would emerge.  But I confess to being surprised by the 
rapidity with which the complications emerged.

One other related comment: as such non-university non-lab sites were 
connected and as the nature of their "usage" became, ahem, mixed, we all 
realized that our options for doing the right thing were limited by the 
essentially-pure destination-IP-address-based routing of the Internet. 
Policy-oriented folks encouraged the IETF and others to explore "policy 
routing", but there was a very natural mismatch between their visions 
and what was practical, then or now.

This is, of course, similar to many human endeavors where practice 
precedes theory.

	-- Guy

On 12/6/15 1:29 PM, Dave Crocker wrote:
> On 12/4/2015 9:26 AM, Jonathan Grudin wrote:
>> I visited USC/ISI in 1993 and remarked that commercial traffic seemed to have begun and was told emphatically that it was because enforcement was lax.
>
>
> Around 1988/89 while I was at Digital Equipment, the Support Services
> folk who were funding my lab said they needed to be able to do regular
> customer service -- ie, service their commercial accounts -- over the
> Arpanet, but couldn't because of the AUP.
>
> However since they saw DEC's competitors already operating this way,
> they said they needed to get the DEC attorneys to find a way to say it
> was ok.  This took some effort, but eventually the lead Corporate
> attorney did bless such conduct.
>
> So, yeah, it was lax.  For a long time.
>
> d/
>