DUMPIT: Fri Dec 2 15:26:42 EST 1983 fav { prss

From: me
Date: #1 Nov 1983 at 1403-EST

To: jsl Comfen Ruf
Subject: Letter to NSF
wert Forwarded Message q

2From: cunyymsmai ler .

Date. 20 Nev 1983 at 1623-EST

Date: 20 November 83 16:20 EST

>From: ITHFCU @ CUNYYM

To: MEL@ROCKVAX

Subject. Science network letter
a,

Dr Edward Knapp

Director

National Science Foundation
Washington, DBC

Dear Dr. Knapp:

The past fifteen years have seen the rapid development of computer
network technology. Today, networks provide a high level of
communications services to many university, government, and industrial
computer scientists aS well as to numbers of scientists and engineers
in other fields. Arpanet (DCA and DARPA sponsored) » CSNET (NSF
sponsored), GITNET (university computer center sponsored), and Mailnet
(Educom sponsored) are examples of communications networks which
currently serve the university community. The very significant
benefits of these networks have been well demonstrated to their users
and sponsors

We understand that NSF is now considering establishment of a ‘Science
Net’ to provide communications services, including supercomputer
access, to scientists and engineers throughout the U.S. university
community. We are writing for two reasons

(1) First, we strongly encourage the NSF to proceed as quickly as
possible to establish a national Science Net. We believe that the
success of the supercomputer initiative will, to a great extent,
depend on the quality of access provided and the availability of
enhanced communications throughout the research community

(2) Secund, we encourage use of existing technologies which we believe
are adequate to deliver required services. Computer communication
networks such as Arpanet, CSNET, and BITNET have established
extraordinary levels of intra and interdisciplinary communication,
a5 well as increased productivity through remote computer access,
software exchange, etc., all to the great benefit of existing users
of these networks.

An example af an important service currently available via
Arpanet/CSNET is the MOSIS fabrication facility sponsored gaintly by
DARPA and NSF. Besides representing an important precedent with respect
to cooperation between the two agencies, MOSIS has demonstrated the
utility of the type of network recommended for providing services to a
widely dispersed heterogeneous user community.

A possible scenario would be to utilize DARPA Internet technology where
the cost of such network connections can be justified by the local

cs

ppee
“"“YaeveT oF research "activity. DARPA Internet technolegy provides tne
highest level of service amongst the currently available connection
alternatives. It is likely that fewer than fifty such nodes would be
Tequired Universities with a lower level of research activity could
access Science Net via use of lower cost remote access to one of these
nodes or by use of one of the CSNET/BITNET technology alternatives
(telephone based relaying. X.25 net connection, leased line).
Interoperability between the Arpanet, CSNET, and BITNET technology
components would allow all these approaches to operate as one
consistent network

Very high bandwidth requirements (greater that 56 kbps) could be
satisfied by adoption of wideband technology under development within
the DARPA research community. We recommend that the NSF cooperate
with DARPA to continue development of this technology and to provide
for establishment of a prototype network for use by the science
community

Given the above approach, the total annual cost of including relevant
U.S. university science and engineering groups would be insignificant
in comparison to the potential benefits. For example: the cost of an
Arpanet connection for the entire scientific and engineering community
Gf a major Tesearch university is a small percentage of the yearly
costs of internal computing services within that university. Most of
the annual cost of such a network would be for the leasing of
communications lines and payment of any associated traffic charges.
Since there would be minimal investment in equipment, such a network
would not be bound to a specific generation of technology and could be
upgraded and/or supplemented by commercial networks when and if they
beccme widely available at an acceptable cost and level of service.

The individuals signing below recommend that you. as Director of the
National Science Foundation, unde: take to supply network services to
the university science and engineering community as soon as possible

Signed by Identification:

(Those marked with an asterisk (*) are either computer scientists
active in the networking field or scientists with first-hand experience
in the use of one or more of the existing networks. )

crete ~ End of Forwarded Message
This letter is being circulated for signatures. Iwill be signing.

If you haven’t already been solicited, we would welcome the contribution
of your name.