[mailhist-discuss] Segments of email history
fuhn at pobox.com
Tue Apr 10 10:36:26 PDT 2012
>On 4/10/2012 7:34 AM, Craig Partridge wrote:
>> I think we're miscommunicating, so let me try again.
>> DNS made it possible to adopt hierarchical names in the Internet. It also
>> provided a possible template for hierarchical naming elsewhere.
>> The decision by other networks to adopt not just hierarchy but the DNS
>> hierarchy, and for the Internet folks to respect subdomains created by
>> other networks, made it seamless -- and that was done at a meeting chaired
>> by Jake Feinler at SRI in January 1986. The other networks had to develop
>> their own lookup mechanisms internally.
do you recall if this meeting included attendance by cc:mail folks?
While there were eventually gateways to smtp from cc:mail, the naming
conventions in cc:mail supported use of spaces. At Intel at this
time (approx 1986-87), most of the business side of the corp was
using cc:mail with names including spaces. Those folks could not be
reached by mail from the Internet. cc:mail also seemed to have had a
deleterious effect on network architecture, especially for a
world-wide organization. Imagine a world-wide bridged network with
odd instabilities caused by the mail system. The first router for
general purpose use came into Intel for use in upgrading the CSnet
connection from PhoneNet.
>> It was entirely possible, and indeed, gently threatened at that meeting
>> that the networks would adopt some other hierarchical scheme.
>Well, this definitely underscores the importance of operational choices, as
>opposed to technology "invention".
>I always highlight a critical contribution of NSFNet as forcing support for
>multiple IP backbones (resulting in the creation of BGP). Prior to NSFNet,
>there was officially only the BBN backbone, with tidbits of other unofficial
>backbones, but they were supported in an entirely ad hoc manner.
>NSFNet forced a
>While I knew that CSNet was the first time Arpa delegated authority for adding
>new sites to the net (albeit through indirect connections, initially) I hadn't
>thought about the emerging set of independent email services as composing
>multiple backbones that were being integrated. You reference to "threat"
>highlights the reality of this.
After a period of dual mail system use, some senior engr. mgs at
Intel pressed the IT folks who were implementing cc:mail on why it
was easier and more reliable to send email half way around the work
on the Internet than to send email to another corporate site via
cc:mail. Encounters such as this led to a large internal effort to
decide on a single network architecture and move forward to
standardizing on it. Internet architecture won out (approx 1992).
These experiences also helped to facilitate the inclusion of native
TCP/IP support in the PC, rather than having to purchase 3rd party
hardware and software.
>What 1-3 word tag would you use for this 'consortium' decision in
> Dave Crocker
> Brandenburg InternetWorking
>mailhist-discuss mailing list
>mailhist-discuss at emailhistory.org
More information about the mailhist-discuss