IP: CPSR comment on ACTA petition concerning "Internet

[Dave Farber <farber () central cis upenn edu>]

Computer Professionals for Social Responsibility (CPSR) is a
15-year-old grass-roots organization of 1,600 members who work in
computer-oriented and related technical fields.  Our mission is to
educate the public, the media, lawmakers, and policy-makers about the
effects of computers and networks on society and the implications for
social policy.  We are a nation-wide organization with many
international members.  One of our recent campaigns has been to
represent the needs of the public in the development of the National
Information Infrastructure.  This document represents the position of
CPSR on the ACTA Internet Phone Petition.


Summary
-------


Viewed superficially, the use of real-time audio transmissions to
carry on bi-directional voice conversations on the Internet is simply
another way to provide telephone service.  This view is the basis for
the ACTA petition.  Our contribution here is to look more closely at
the technology of real-time audio and to offer a more technically
accurate view of it as software and as a component of emerging
Internet applications.


In this comment, we take a broad view of Internet applications and
draw long-term conclusions about the effect that regulation or an
injunction would have on research and development in the field of
communications.  Like the Congress and the FCC, we look forward to
competition in new technologies that can lead to eventual improvements
to the public in service and cost.  The constellation of technologies
that make up Internet telephone (and the related innovations discussed
in this comment) promises just such improvements.  Attempts to
regulate an isolated application on the Internet, as we shall see,
could kill the goose that lays the golden eggs.


The main points we make are:


* Real-time audio on the Internet is essentially just one component of
  the many useful applications being developed in digital networking.


* Real-time audio is too different from telephone service, both
  technically and in terms of quality, to apply regulations related to
  telephony.


* Internet telephone products are software, which are not regulated by
  the FCC.


* Attempts to single out audio and regulate Internet telephone could
  prove harmful to technological progress in digital networking.
  Allowed to flourish, this software could drive the public's demand
  for better underlying Internet service.


* Regulation would run into insurmountable technical barriers.


* If Internet telephone grows to the point where its use has a
  negative effect on funding for the local telephone infrastructure, a
  restructuring of the pricing structure should be found that takes
  technological evolution into account, rather than suppress it.


The Place of Audio in Emerging Internet Technologies
----------------------------------------------------


Increasingly, digital electronic networks are being used to carry
graphical images, audio signals, and video signals.  Once they are
sampled and represented in numerical form, audio and video become just
another form of digital data and can thus be carried over a digital
network like any other traffic.


A large number of organizations are working hard to make the Internet
more than a text-only medium.  Many people see the availability of
graphics, audio, and video as key to the broadening of digital
networks as valuable media for education and other social goals.
Voice transmission has been used for such experiments as Internet Talk
Radio (in which files of audio data can be downloaded and played
off-line) and more recent experiments in real-time radio transmission.
CU-Seeme has been available for several years to provide audio and
video links between individuals on the Internet.  While the ACTA
petition considers Internet telephone a "new technology," we recognize
it as a convergence of many existing techniques that grow naturally from 
Internet applications and audio sampling.


One of the most exciting trends in Internet technology is the
combination of text, audio, and video in real-time collaborative work.
Here, all data is transmitted as a single stream and then divided into
constituent channels at the receiving end.  We mention these
collaborative multi-media applications because they illustrate the
seamlessness of digital technologies.  There is no technical basis for
distinguishing between audio and other media.  In all the digital
networking technologies currently envisioned for the near future (such
as Asynchronous Transfer Mode) transmissions are still divided into
packets, and many kinds of data can be combined and intermingled.


Technical Implementation and Limitations of Real-Time Audio on the Internet
---------------------------------------------------------------------------


The products known as Internet telephone require equipment for
recording and generating sound, plus software to transform it into
transmission units acceptable as Internet traffic.  The sender's
computer samples the sound generated by the sender and digitizes it.
After compression, the sound samples are broken into transmission
units that can be handled by lower levels of the Internet protocols.
These packets are sent out on the Internet bearing the address of the
receiver's computer system and are multiplexed with other traffic.
The medium for the traffic may be a local network or any other medium
that can carry electronic signals, but most users at some point
utilize telephone lines to cover long distances.  The receiver's
computer collects and reassembles the packets, passing them up to
higher layers that decompress the data and perform digital to analog
conversion.  The resulting signals are then relayed to speakers.


It is important to understand that Internet telephone is not a
service.  It is a collection of software components that users install
on their computers.  They can then transmit sound as often and as long
as they want.  Furthermore, a wide range of communication channels can
be used for Internet traffic.  While some have suggested that
real-time audio can be subjected to long-distance tariffs because most
users employ telephone lines at some point in their connection, the
use of these lines is really incidental.  Any other medium that can
carry Internet traffic would has the same utility for real-time audio.
Internet telephone, in short, is a form of software totally different from 
the services traditionally regulated by the FCC, and is not
suited to the regulatory framework.


While Internet telephone is beginning to find commercial use, most
real-time applications remain experimental because sound and image
quality remain low for the vast majority of Internet users.  The
problems include:


1. Slow transmission speed for most users.  The bandwidth of each
   user's Internet connection determines how many digitized bits can
   be transmitted at one time, and therefore the accuracy with which
   sound is reproduced.  In most cases sound quality is not as good as
   telephones, although this could be remedied in the future as more
   users demand higher bandwidth and the demand causes costs to
   decrease.


2. Half-duplex connections.  Many products allow only one side to
   transmit at a time.


3. Unreliability.  Because all Internet traffic is broken into packets
   and sent over a variety of channels rather than a dedicated
   circuit, users can experience short delays.


If allowed to proceed, however, experiments in multimedia can lead to
improvements in service through either physical upgrades or new
protocols.  We caution against interfering with demand for
technologies that can lead to improvements in service over the long
run.  Regulating voice transmission on the Internet could be
technological infanticide.


Implications of Singling Out Real-Time Audio for Regulation
-----------------------------------------------------------


Given that audio becomes just another form of packet data on the
Internet, an attempt to halt or regulate its use would lead to
technical dilemmas and unintended effects on technological progress.
Internet phone technology cannot be banned, because several free
products are already available and the only way to eliminate their use
would be to exhaustively check the hard drive on every American's
computer.  Any restrictions placed on U.S. manufacturers would simply
cause them to move overseas.  (In fact, one of the first companies to
offer phone products for the Internet is located in a foreign
country.) Furthermore, a ban would fly in the face of the basic
principles that drive technological development: the promotion of
competition, the desire for improved technology, and the goal of
serving end-users better.


Special treatment of audio data, even if it were feasible, would place
barriers in the way of technological development.  A central premise
of digital network research is that all data can be treated as
fungible (that is, any type can be replaced with any other without
changing the behavior of the network software).  The protocols assume
that all data will be reduced to a lowest common denominator and be
freely combined with other forms of data.  The most advanced and
promising technologies, as stated earlier, mix several media.  The
technical challenges of trying to measure and price one of these data
streams may render the technologies unusable.  It is crucial,
therefore, to the development of this technological frontier that
audio be treated equally and not be subject to special regulation.


Inapplicability of the Telephone Model
--------------------------------------


While useful as a substitute for telephone use in some situations,
real-time audio should not be compared to telephone service for
several reasons.


1. It is not a service.  It is a software application that runs on
   users' computers and utilizes networking protocols.


2. It operates in an entirely different manner from telephone service,
   offering packet-based transmission instead of a dedicated
   circuit-switched connection.  Under current conditions, this
   produces the differences in quality mentioned earlier in these
   comments.


3. It can serve as one component of multimedia applications, thus
   giving rise to conditions where the audio portion of transmission
   cannot be separated from other Internet use.


4. Different products use different protocols, so the owner of one
   product cannot necessarily communicate with the owner of another.


The important criterion for establishing rules is to preserve the
viability and robustness of the medium.  The ACTA petition, as we show
in this comment, cannot be granted without imposing serious burdens on
this promising area of development.


Equity in Pricing
-----------------


CPSR certainly subscribes to the principles that funding of the
telephone infrastructure must be assured, and that pricing of services
should be equitable.  But the use of the Internet for audio
conversations is still an emerging technology, and due to the many
changes may take place technically and financially, it is inadvisable
to set up regulations now.


Current quality is hampered by bandwidth that is slightly too low for
most users and unreliable delivery of packets.  If the market for
Internet telephony is encouraged to grow, these problems may be
remedied by the installation of new hardware or the use of protocols
that promise a higher quality of service.  When the market adjusts in
this manner, it may be accompanied by new pricing structures.


In any case, the public is not well served by placing barriers in the
way of new technologies.  Internet telephone is currently used too
sparsely to have an impact on funding for the telecommunications
infrastructure.  Should it grow in popularity to the point where it
has an impact, the entire pricing structure should be reviewed to
determine where funds should come from.  The main cost benefit of
using Internet telephone, however, does not stem from avoiding fees
but its use of the Internet for sending traffic.


Future Potential
----------------


Telecommunications industries should, in our view, encourage and
improve the conditions for promising technologies that could reduce
costs or provide new functions.  Real-time audio transmission over
digital networks offers intriguing potential benefits: lower costs,
user options such as storing and filtering, and integration into
collaborative multimedia.  The ACTA position goes precisely in the
wrong direction.  Instead of furthering the potential of new
technology, it puts barriers in the way.


For the reasons stated in this comment, CPSR recommends that the
petition be rejected and that the FCC take no action related to
Internet telephone at this time.


Andrew Oram, for CPSR
<andyo () ora com>