[1610] in Commercialization & Privatization of the Internet
Internet articles
daemon@ATHENA.MIT.EDU (Telecom-Editor)
Wed Dec 4 16:18:17 1991
Date: Wed, 4 Dec 91 20:46 GMT
From: Telecom-Editor <0003101692@mcimail.com>
To: "com-priv@psi.com" <com-priv@psi.com>
It appears that the ftp arrangement I mentioned earlier will take
somewhat longer than originally anticipated to put into place (still
working on it, however). In the interest of timeliness, I'm going to post
the two articles here, starting with Jay Habegger's and then followed by
Vint's in a separate posting (they should be read in sequence). Fair
warning: Jay's is quite a long piece. Also, the articles are "text only"
so if you want to see the graphics, you'll have to pick up a copy of
Telecommunications :-) Thanks again for the interest.
Tom Valovic
Editor
text follows:
WHY IS THE NREN PROPOSAL SO COMPLICATED?
Jay Habegger
Editor's Note: The following article is based on a paper that was one of
the finalists in the recent ICA student paper competition. It presents an
interesting but controversial view on the impacts of NREN legislation on
national telecommunications policy. -- TV
A national gigabit network has moved one step closer to reality with the
passage of the High Performance Computing and National Research and
Education Network Act of 1991 by the Senate on September 11, 1991. A
similar version of the bill was passed by the House in July. The
differences between the two bills must be resolved in conference, but
passage of a bill is almost guaranteed. Placing aside initial reluctance,
the Bush administration has also included support for many measures of
the bill in its FY 1992 budget request.
One component of both the Senate and House bills is the National Research
and Education Network (NREN). The legislation calls for the construction
of a high-speed computer network to link academic, industry, and
government agencies throughout the US. While early legislation
specifically called for the implementation of a 3-gigabit network, the
latest language only mandates a multigigabit network by 1996. The sponsor
of the Senate bill, Senator Albert Gore, has declared that computer
networks, such as the NREN, are to become the interstate highways of the
information age.
Despite its popularity among lawmakers and widespread support from
academic, industry, and government agencies, there has been little
analysis of how the NREN is going to emerge and the impacts of its
construction. When the rhetoric surrounding the proposal is removed,
surprisingly few details can be found anywhere about what the network
will look like, which constituencies it will serve, the role for the
private sector, or how the NREN will affect US telecommunications policy.
A NEW NATIONAL NETWORK
The NREN proposal was inspired by the Internet, the collection of
computer networks that presently reaches many institutions of higher
education, government agencies, and a significant number of industrial
locations in the US. The Internet also includes networks in Western
Europe and the Pacific Rim. A recent estimate places the number of
individuals reachable on the Internet at around 3 million.
The US Internet is a heterogeneous collection of computer networks
organized in a hierarchy as depicted in Figure 1. Local area networks
compose the lowest level of this hierarchy and are typically connected to
a midlevel network, which is in turn connected to a backbone. The largest
backbone is the NSFNET, which is partially supported by the National
Science Foundation (NSF). Other national backbone networks are operated
by Performance Systems International (PSI) of Reston, VA, Advanced
Network & Services (ANS) of Elmsford, NY, and UUNet Communications
Services of Falls Church, VA. ANS also operates the NSFNET network under
subcontract from Merit Inc., the organization that operates the regional
midlevel in Michigan. Merit has a cooperative agreement with NSF to
provide the NSFNET backbone.
The increasingly popular TCP/IP protocol suite is used on networks
directly connected to the Internet and permits a high degree of
interoperability between the disparate networks attached to the network.
In addition to the interoperability provided by TCP/IP, there are also
gateways to other networks, such as MCI's electronic mail service, X.25
public networks, and dialup-based networks, such as USENET and FIDONET,
which create a pervasive "matrix" of computer networks that reaches an
enormous number of individuals beyond those reachable on the Internet
alone. These gateways to other networks typically provide limited access
to the Internet; only electronic mail can usually pass between networks.
NREN ARCHITECTURES
The networks that compose the matrix can not only be characterized by
their design philosophy of modest speeds and functionality, but also by
wide connectivity. Even the Internet -- by far the fastest network in the
matrix -- has wide area connections that typically run at 56-kbps or T1
rates. A project to upgrade the Internet backbone to T3 rates is
presently underway, but not yet completed (about one-half of the T3
network is operational). The most basic level of functionality offered by
networks in the matrix is electronic mail. Again, the Internet offers the
greatest functionality including electronic mail, file transfers, and
remote access to computing resources. Figure 2 shows the breakdown of
Internet traffic by application, and Figure 3 shows the growth in traffic
on the NSFNET backbone. The figure only captures traffic that crosses the
backbone, and is therefore not an absolute measure of Internet usage
since traffic confined to the midlevels or campus and corporate local
area networks is not included.
Discussions surrounding the NREN hint at a radically different
architecture. The NREN is to be a multigigabit network offering 1000
times the capacity of the Internet T1 backbone and a plethora of advanced
telecommunications services. Many members of the science community have
advanced the idea that the NREN is going to provide gigabits to the
desktop so they may run applications, such as scientific visualization,
which involves displaying entire pixel arrays in a movie-like fashion.
There is also considerable interest in developing a national digital
library, which many believe will one day hold the complete digitalized
works of the library of Congress if the technical and legal issues can be
resolved. The NREN would provide access to this library.
The scope of gigabit access is an issue without easy answers. Economics
dictate that the gigabit portions of the network are going to remain
confined to the backbone for some time; even T3 is not yet tariffed in
many portions of the US. On the other hand, Senator Gore seems to support
the vision of the NREN as a network with gigabit pipes everywhere. He has
also introduced a trial balloon piece of legislation that would free the
RBOCs from the line-of-business restrictions of the MFJ in exchange for
the RBOCs' implementation of a plan to install fiber optic cable in most
subscriber loops by 2015. Some supporters of the NREN have characterized
it as the first step toward a network that would reach into every home
and business.
Federally funded research into gigabit networks is already proceeding.
Not all of this is research into the nuts and bolts of gigabit
networking. Some of it involves developing gigabit applications. The
Corporation for National Research Initiatives (CNRI) in Reston, VA,
administers a 3-year, $15.8 million grant from NSF and the Defense
Advanced Research Projects Agency (DARPA), in addition to in-kind support
contributed by the industry, which is being used to fund a group of
gigabit test bed projects. The projects typically involve cooperation
between universities, government labs, and industry participants. For
example, one task of the VISTANET test bed, which includes participation
by Bell South, GTE, MCNC, and the University of North Carolina at Chapel
Hill, is to develop applications for gigabit networks.
Many of the constituencies that have supported the NREN are presently
interested in applications that require only modest data rates, but wide
connectivity. Distribution of information through file transfer, and
communication via electronic mail is what the Internet is mostly used for
today. Supporters of an NREN architecture that places an emphasis on wide
connectivity would like to see the availability of these applications
expanded. The library community would like to implement a nationwide
bibliographic data base. The educational community would like to advance
from the dialup access typically available to them in order to make use
of distributed expertise and make information available to students.
A detailed NREN architecture plan has yet to be developed beyond the idea
that the NREN will evolve from the present Internet in three stages. One
can be sure that the battle is just beginning over a network architecture
that emphasizes high data rates everywhere versus one that emphasizes
widespread connectivity. The Senate version of the NREN bill authorizes
about $1 billion over the next 5 years divided between NSF, NIST, and
NASA to develop the network, and carry out significant research in
high-performance computing. Only a fraction of this is actually earmarked
for the NREN. NSF has allocated $195 million over 5 years. Funds for the
Department of Energy (DOE) and DARPA are covered under separate
authorizations. This sum is not likely to be enough to build a network
that will satisfy both sets of constituencies, and given the budget
situation, it appears equally unlikely that funding for the NREN will
increase significantly.
PRIVATE SECTOR ROLES
As the Internet has increased in speed and sophistication within the
academic community, personal computers and LANs have become more
prevalent in every sector of industry. Computing prices that continue to
fall, coupled with the recognition of the usefulness of LANs, can only
cause an acceleration in the number of LAN installations. Since LANs
represent the lowest level in the Internet hierarchy, the increasing
number of LAN installations creates a pool of demand for access to
services available through the Internet. As mentioned before, the most
widely used of these services are distribution of information by file
transfer and electronic mail, but one can easily imagine other uses in
the corporate world, such as electronic data interchange (EDI) between
firms and access to commercial data bases (such as LEXIS/NEXIS or
Dialog). Note that these are also applications that place an emphasis on
widespread connectivity between different firms. The economics of both
data and voice networks are such that private networks are likely to
remain the solution of choice for significant amounts of traffic within a
firm, but a network with widespread connectivity throughout industry
could also be used for traffic among firms.
The NREN includes the words "research" and "education" in its name, but
it is also supposed to link industry. Because of this, many US
telecommunications firms view the NREN as a program for the US government
to shoulder the risk in developing new markets and leave the profit
making to private firms. According to Robert W. Lucky of Bell
Laboratories in testimony before Congress, "the business people do not
think there is money in it right now. But there will be, and there can be
a tremendous market if you [the government] get out and stimulate it for
us." O. Gene Gabbard, then chairman and CEO of Telecom*USA, suggested
that the government should establish a nonprofit corporation to build the
network, and then the network should "be transferred back to commercial
operation within 2 years thereafter."
Language in the NREN legislation states that the network should "be
phased into commercial operation as commercial networks can meet the
networking needs of American researchers and educators," but what this
passage means is far from clear. Lawmakers do not really fully understand
what this means either, and have called for a report on the
commercialization of the network to be prepared. The report is due 1 year
after the NREN program gets underway.
With the number of constituencies lining up for a piece of the NREN pie
and all of these constituencies expecting either free or nominal access,
phasing the NREN into commercial service promises to be a difficult task.
The question immediately arises as to how a viable private competitor is
to emerge in the presence of a sophisticated subsidized public network?
The private firms that are attempting to provide an alternative to the
publicly funded NSFNET backbone already face difficulties because NSF is
the purchasing agent for almost all of the potential customers. NSF
purchases service from ANS through a subcontract from Merit Inc. Because
NSF acts as a purchasing agent for most of the customers for Internet
services, it has created a monopsony market where it is by far the
largest buyer.
If development of the NREN proceeds in the same fashion, the result is
likely to be even worse. Commercial traffic is presently prohibited on
the NSFNET backbone by "acceptable use" policies that require traffic
crossing the backbone to be consistent with the research and education
purposes for which the network was intended. These restrictions force
some users of the Internet to choose a service provider in the open
market. In contrast, the language of the NREN legislation indicates that
many forms of commercial traffic are going to be allowed on the NREN.
Disturbing to some, the NREN legislation states that the "use of the
network by commercial information service providers" should be
encouraged. This has the potential to short circuit the fledgling efforts
of private firms, such as PSI, to offer Internet-type services.
A NEW TELECOMMUNICATIONS POLICY?
The President's science advisor, Dr. Allan Bromley, stated to Congress
that he "would like to see us planning for a system that somewhere in the
not too distant future would have the same characteristics as the
telephone system." This national network that reaches as far as the
telephone network is often referred to as the "National Network" (NN).
The NREN is seen as the first step to this network. Bromley continued,
stating that the NN "would be a utility, hopefully by that time supported
in the private sector, that would make broadband communication available
to the individual home and to the individual user [emphasis added]."
There are two possible paths to the NN from the NREN. First, as implied
by Bromley and stated by others, the NREN could serve as a demonstration
project to show what can be done with high-speed networks, and the
private sector can construct the rest of the NN around the NREN. The
second, and much less discussed path, is that Bromley's hopes are not
fulfilled and that the government operated NREN evolves directly into a
government operated NN.
Development of a ubiquitous computer network offering a plethora of
services by the public sector challenges established US
telecommunications policy. If the second path is chosen and the
government does not leave the field as a result of constituency pressure,
or because the private market was quashed by a subsidized network, the US
government would become a telecommunications service provider, posing an
even greater challenge to US telecommunications policy. According to Dr.
Craig Fields, director of DARPA in 1989 when he testified to Congress and
a supporter of the NREN, the legislation suggests a "rather active role
for the government -- a more active role."
The Internet has developed outside of the telecommunications industry,
although transmission facilities have always been provided by the
carriers. ARPANET, precursor of the Internet, was originally conceived as
an experiment in fault-tolerant communications for the Defense
Department. NSF started developing NSFNET in 1986 as a method of
providing investigators with remote access to supercomputer facilities.
Use of the Internet has remained largely the province of the academic
community ever since. It is, therefore, not surprising that the entities
lining up for involvement in the NREN are none of the conventional
players in US telecommunications policy.
NSF, DARPA, and DOE are the lead agencies in implementing the NREN.
Traditional players in telecommunications policy, such as the FCC and the
Department of Commerce, have been notably absent. It is also interesting
to note that the subcommittees that handle telecommunications issues in
both houses of Congress have had nothing to do with the NREN proposal.
According to one Capitol Hill staffer, the NREN proposal "shoots a hole
in the FCC's jurisdiction a mile wide."
The US government had a brief role in the development of the telegraph,
and during the early history of US broadcasting, it was suggested that
the government should operate all radio stations. The path of government
development and ownership of any telecommunications medium was rejected,
of course. Chosen instead was a system of private development and
operation with regulatory oversight administered by the FCC. The public
entities involved in the development of the NREN can be entirely
different from those that regulate other telecommunications mediums
because NSF, DARPA, and DOE do not have instructions from Congress
defining their role, as the 1934 Communications Act did for the FCC,
because they have new constituencies, and because they are unburdened by
the administrative framework that has evolved at the FCC.
The US telecommunications industry has embraced government development of
the NREN and remained silent on the precedents it established. In all of
the hearings before Congress, only Tracey Gray of US Sprint expressed
concern about the language in the NREN legislation. According to Gray,
the Congress is creating through the NREN a "telecommunications
infrastructure that parallels the commercial, public facilities with a
government-owned and operated telecommunications system."
Jay Habegger is pursuing an MS in telecommunications at the University of
Colorado at Boulder, and he also holds a bachelor's degree in
Mathematics. His master's thesis is a study of the NREN proposal. He can
be reached on the Internet at habegger j@cubldr.colorado.edu.
***
This article first appeared in the November issue of Telecommunications
Copyright 1991
Telecommunications Magazine
