[387] in Commercialization & Privatization of the Internet
what could have been
daemon@ATHENA.MIT.EDU (Dennis Perry)
Fri Mar 15 14:25:36 1991
Date: Fri, 15 Mar 91 14:12:34 EST
From: perry@MCL.Unisys.COM (Dennis Perry)
To: com-priv@psi.com
Cc: perry@mcl.unisys.com
>From perry Fri Mar 15 14:11:07 1991
Received: by kauai.MCL.Unisys.COM (4.1/mls/3.3)
id AB05921; Fri, 15 Mar 91 14:10:33 EST
Date: Fri, 15 Mar 91 14:10:33 EST
>From: MAILER-DAEMON (Mail Delivery Subsystem)
Subject: Returned mail: User unknown
Message-Id: <9103151910.AB05921@kauai.MCL.Unisys.COM>
To: perry
Status: R
----- Transcript of session follows -----
Connected to psi.com:
>>> RCPT To:<com-pric@psi.com>
<<< 550 <com-pric@psi.com>... User unknown: Operation would block
550 com-pric@psi.com... User unknown
----- Unsent message follows -----
Received: by kauai.MCL.Unisys.COM (4.1/mls/3.3)
id AA05919; Fri, 15 Mar 91 14:10:33 EST
Date: Fri, 15 Mar 91 14:10:33 EST
>From: perry (Dennis Perry)
Message-Id: <9103151910.AA05919@kauai.MCL.Unisys.COM>
To: nren-discuss@psi.com, rma@tsar.cc.rochester.edu
Subject: Re: Atacking the regional networks
Cc: com-pric@psi.com, perry@mcl.unisys.com
Richard writes about the "much maligned regional networks" from what
I would call a "narrow" perspective. To say that if the regional's
"hadn't accepted the networking challenge or opportunity which the
NSF supercomputer program presented us with in 1985 there would be no
1991 Internet as we know it." is probably a true statement, but perhaps
not in the way Richard ment it to be. I would argu that mess we are in
now is a direct response to the formation of regionals and the development
of what I feel is a flawed model that is being used to develop the Internet.
Now, we all have our biases, and I am no exception. My bias comes from
being at DARPA during the years of late 1985 to late 1987 and responsible
for the problems, solutions, etc. that the Arpanet was going thru at that
time. One of those problems was working with, initially Denis Jennings and
then Steve Wolff, to expand the Arpanet to accomodate the initial
NSF networking requirements. DARPA and NSF developed and signed an MOU
to do certain things together and in response to each others' needs.
As an aside, the MOU was not adhered to in any substantial manner. But,
the NSF decided that they wanted to do things a certain way, i.e. develop
regional networks and a hierarchy of connectivity rather than expand the
Arpanet thru a well engineered approach to accomodate growth. DARPA, on the
other hand was beginning to feel that the operation of a network ( at least
the cost of that network) was taking too much away from the funding of
research into networking. As I recall, we were paying about $10M operation
costs for lines, etc. When I left DARPA, the dismanteling of the Arpanet
begin in earnest, the NSF was left with the opportunity to push their
architecture. There were other options, presented to DARPA in the form
of proposals by some of the major players you still see in the NSFnet world,
especially those involved in higerspeed networks.
To make it short, I agree what we have is not what we could have had,
we could have had it better sooner!
dennis
