|
|
June 1994
The Science, Technology
& Public Policy Program of the John F. Kennedy School of Government
and The National Institute for Standards and Technology
STANDARDS DEVELOPMENT
AND INFORMATION INFRASTRUCTURE WORKSHOP
Today's Cooperative
Competitive Standards Environment For Open Information and Telecommunication
Networks and the Internet Standards-Making Model
by
Today's Standards
Making Architecture
The architecture of standards-making
organizations in the telecommunication and information fields has undergone
fundamental change over the past decade. The old architecture was simple
and well-bounded around a handful of bodies with explicit internation al,
regional, national, and subject matter jurisdictions. These standards-making
bodies were virtual sovereign, following slow, deliberate, time-honored
processes that remained essentially unchanged for the preceding 130 years
since the first multilateral telecom standards conference, and engaged legions
of standards professionals whose careers often began and ended in a single
committee.
Over the past ten
years, that old architecture has been fundamentally altered. Constellations
of new bodies now exist with diverse new constituencies and boundaries,
and all are competing in a global standards marketplace. Even the form
of these new bod ies differs dramatically from traditional organizations.
Their range includes:
- industry aggregations
around a vendor specification,
- ad hoc global initiatives
around a specific technology,
- national or regional
bodies created to bring about a competitive marketplace,
- global hyperdynamic
developmental and technology transfer "engines" like the Internet Engineering
Task Force.
Meanwhile, the traditional
bodies struggle to evolve within a standards marketplace that finds their
products largely unacceptable, yet still running processes that incur collective
costs of tens of millions of dollars per year.
This transition hasnít
been simple or easy. The notion of competition in standards making-like
competition in a rigid monopoly provisioning environment or socialist
economy-is not accomplished without considerable angst and difficult accommodation
by tho se relinquishing power centers and jobs.
The chart depicting
the Standards Making Universe-popularly known as the
Rosetta Stone seeks to provide perspective and portray relationships
within this new architecture. The chart was originally prepared for the
first Standards Summit in 1990 and has been revised continually since
that time to reflect the changing architecture.
Why is this rapid
transition occuring?
The reasons fall into
several categories:
- Moore's Law (i.e.,
electronic technologies are changing dramatically on an average of every
two years). Furthermore, in the highly dynamic environment of the Internet,
fundamental rates of change measured in months. (Rutkowski's Law)
- Most telecommunication
and information markets are very competitive. The marketplace, not institutions
and government, decides winners and losers. The most classic current
example is Open Systems Interconnection (OSI). The publisher of Communicatio
ns Week International, at a recent industry forum of CEOs, chided European
Union leaders about "...effectively killing advanced data networks in
Europe through single-minded pursuit of OSI solutions."
- Most of the information
infrastructure has passed from being a public good to now being a private
commodity. Millions of individuals and organizations now own and design
a collective national and global infrastructure. The Interop trade shows,
for e xample, now the largest industry events in the world, are a manifestation
of this transition. They are also an example of a new kind of industry-based
institution that implements interoperable solutions far more effectively
than government mandates.
- An increasingly
global competitive environment effectively precludes solutions favoring
a particular country or market segment. Attempts by governments to mandate
specific directions that are at odds with the global marketplace will
likely only disad vantage that nation or market by limiting both the
quality and performance of available products and services to users,
and the scale of the market available to vendors.
- The requisite manner
in which standards are developed, promulgated, and implemented for computer
network environments is fundamentally different from hardware-oriented
fields. Although not quantifiable, the development of computer software
appears to require a rather different culture-a handy reference that
captures the kinds of individuals, institutions, and processes necessary
for success in this environment.
- Time-to-market
has become the single most compelling factor for both service providers
and product vendors. This concern is a byproduct of rapid technology
change, a robust competitive marketplace, and a globally competitive
environment. Time-to-mar ket encompasses not only rapid development
of standards, but also implementability and meeting real user needs.
- The last twenty
years have been an expensive collective learning experience about "bottom
up" versus "top down" initiatives. Top down initiatives are characterized
by grand telecommunication and information infrastructure standards
programmes begun t hrough traditional international organizations. In
these organizations, long-term concepts and plans are developed after
years of deliberation and then pursued and implemented at regional,
national, and local levels. This process can sometimes take deca des.
Meanwhile, the real revolutions in the telecommunication and information
fields have occurred from the bottom up. Personal computers and workstations,
local area networks, cost-oriented leased lines, routers, network operating
systems, the Internet, and other capabilities have empowered individuals
and organizations to develop their own infrastructures, and control
their own information destiny.
These factors have produced
a very different standards making architecture. Today, direct government
involvement in picking winners and losers is likely to be the kiss of death
for the unlucky recipient. With few exceptions, every direct governmental
in trusion into the standards marketplace over the past decade has had major
adverse consequences. On the other hand, minimal government involvement,
designed primarily to foster research, collaboration and technology transfer
among developers and rapid dis semination of standards, appear to work well.
Stature of Standards
Making Organizations
Recently, many attempts
have been made to aggrandize some organizations and their products by referring
to them as de jure. This term is usually in contrast with other organizations
and their products that they charactere as de facto. It is not cl
ear how this de jure versus de facto notion was started, but
the terms have fairly specific meanings in law that are wholly inapplicable
to our voluntary systems of standards. De jure means legitimate,
just, or imposed as a matter of law. De facto is a contrasting condition
characterized as illegitimate, condoned, or accepted for practical purposes.
In a world of heterogeneous,
voluntary standards making bodies, no organization has a right to claim
its standards are more legitimate or legally binding or even "preeminent"
than those produced by any others, including individual corporations that
have o btained adoption of their standards in an open marketplace. The
ISO, for example, is a private, not a treaty organization. Even the ITU-T
-which is an international body under a public intergovernmental organization-does
not produce legally binding stan dards. Indeed, at decades of formal international
conferences, great care has been taken to assure that standards remain
purely voluntary-on a par with all other organizations.
Internet Standards
Making as a Model
The Internet standards
development process is by far the best in the business. More than just a
standards process, it is a distributed collaboration and innovation engine
that has produced a thriving new field of electronic communication and a
ten-billion dollar global marketplace growing faster than any communications
technology yet devised. Its very uniqueness, however, suggests that it may
not be easily applied to existing standards making organizations and their
proceedings.
It's worth examining
the attributes of the Internet standards and the associated
processes.
- Individual participation.
From the outset, the Internet standards process was based on individual
as opposed to organizational participation. In fact, organizational views
are not introduced or discussed. This significantly alters behavior a
t meetings emphasizes substantive issues.
- Direct open participation
by experts and innovators. Anyone may immediately access all relevant
information and standards, or may participate in any Internet standards-making
activity. This may be done via the global Internet at no cost, or by attending
any of the triannual meetings at nominal cost. These meetings are also
multicasted live on two audio and video channels to more than 500
sites in nearly 20 countries. This exceptional accessibility has proven
a magnet for experts and enthusiastic innov ators, who freely share their
ideas, expertise, and even their computer code. Many students and low-level
researchers-who freely invent, criticize, and produce concepts and products-are
also drawn into the activity. Much of the work itself progresses on the
Internet-day and night. Much of the work itself progresses on the Internet
- day and night.
- Output consists
of demonstrated working standards. Before Internet standards reach
a certain point, at least two independent implementations must have been
completed. This emphasis on working code and demonstrated interoperability
is consider ed central to the process.
- Emphasis on meeting
real user needs. The use of preliminary interest groups to initiate
a standards making activity, combined with participants who actually use
the technology and the development of real implementations, produces products
that generally meet actual user needs. This occurs predominantly through
"bottom up" rather than "top down" standards-making.
- A well managed
development process. Standards-making is closely followed by Area
Chairs and forced to proceed rapidly or face termination.
- Minimum institutional
ossification. Working
groups are created easily and terminated quickly upon completion of
their specific tasks.
This constant turnover prevents permanent committees, rigid institutional
infrastructure, or semi-permanent individual roles.
- Internet standards
must be accepted by both the Internet Engineering Steering Group and the
Internet Architecture Board. This peer consensus is reached by people
who are intimately familiar with the technology and have one principal
motivation -making sure the standard will work. All formal standards actions
are published electronically and on paper by the
Internet Society - which also takes global international organization
responsibility for the s tandards and
peer liaison with other international organizations.
- Standards and
related materials are universally and instantly accessible and browsable.
Internet standards (and frequently the associated code) are distributed
and made available instantly on international Internet servers by mail-based
and ftp services. Recently, the IETF Secretariat has advanced the state-of-the-art
in standards making support by providing
gopher-based and
WWW-Mosaic hypertext browsing capabilities.
- Activities are
network based. Standards-making on the network also involves rather
considerable support requirements. For each Internet Standards meeting,
this support includes constructing a rather substantial enterprise internet,
obtaining s cores of computers, providing docking stations, and assembling
a multicasting facility. However, this allows attendees not only to accomplish
their work, but also continue their personal professional endeavors.
- Creating the right
culture. Having the right institutional ambiance is very important
to attract the best and the brightest in computer programming and networking.
The right ambiance includes informality, network access, and the presence
of a large peer group. Culture is also an occasionally troublesome as
programmers and networkers have low thresholds of tolerance controls and
influences perceived as unnecessary. Nevertheless, culture is often a
critical factor in determining productivity a nd innovation.
The Internet standards
process-although close to an ideal development model-is quite different
from most existing standards making bodies. While it might be possible
to adopt many of these Internet practices for a new organization, it
is quite different to make over existing organizations to assume all
of these attributes.
Standards bodies
are more often homes for specialized industry or government constituents
than they are neutral technological forums. As a result, even purportedly
open governmental standards forums are usually effectively closed with
no incentives to ad mit outsiders. All of these factors limit propagation
of the Internet model-even though its adoption would clearly be beneficial.
The Government
Role
The appropriate successful
role of government in standards-making is one which encourages generic
open information systems platforms and processes, promotes open technology
transfers among the broadest possible range of innovators, developers,
and users, and allows a robust competitive marketplace to determine winners
and losers. An appropriate international role is to assure that these
same values are applied to multilateral and national forums.
How government
accomplishes these goals is critical to their success. Fortunately,
we have nearly 20 years of benchmarks to gauge what works and what doesnít.
Avoiding picking winners and losers extends to forums as well-although
providing support to op en up the processes does seem to produce significant
benefits.
Perhaps one of
the principal roles of government in this environment is simply to follow
and understand what is occurring both domestically and worldwide. This
information can be made publicly available and used to enhance another
important role-effectin g the open ìtechnology transferî
noted above. A great deal can be done to encourage more open standards
processes throughout the world, which will become increasingly important
as a robust global marketplace emerges and WTO trade rules apply to
the info rmation infrastructure.
|