A Layered Approach to
Internet Legal Analysis
Craig McTaggart*
its
the Internet
that reflects
The analysis of Internet legal and policy issues is
aided by an understanding of the Internets unique,
layered architecture. This article proposes a conceptual
model of
layered
architecture. The model is offered to decision-makers,
policy-makers, and legal analysts, not only as a
roadmap or guide to understanding the Internet, but also
as a tool for identifying Internet legal issues with the
appropriate degree of granularity. Precise identification
enables state and legal actors to assess the impact of
policy choices in a comprehensive manner by allowing
them to consider the implications of those policy choices
for the Internets various elements.
Following an introduction to the Internets four
conceptual layers along with its sublayers and elements,
the author offers a typology of representative legal and
policy issues for each layer. A number of issues are
surveyed with regards to each sublayer. The specific
examples of browse-wrap licenses, overlay software, e-
mail service, and Internet protocol telephony are
discussed
the
characteristic features and concerns of each layer of the
Internet.
in detail
in order
illustrate
to
La comprhension de larchitecture unique de
lInternet favorise lanalyse de ses enjeux juridiques et
politiques. Cet article propose un modle conceptuel de
lInternet qui reflte son architecture sous forme de
couches superposes. Ce modle soffre aux dcideurs
politiques et analystes juridiques non seulement titre de
guide leur permettant de comprendre lInternet, mais
galement titre doutil pour identifier de manire
approprie et prcise les enjeux juridiques qui y sont
relis. Une telle identification permet aux acteurs
juridiques et tatiques dvaluer de manire complte et
dtaille limpact des choix de certaines politiques et de
considrer les implications de ces choix pour les
diffrents lments de lInternet.
Suite une introduction aux quatre couches
conceptuelles de lInternet, de mme qu ses sous-
couches et autres lments, lauteur nous offre une
typologie des enjeux juridiques et politiques associs
chacune de ces couches et sous-couches. Les
exemples des browse-wrap licenses, des overlay
software, des services de courriels et de la tlphonie IP
sont discuts en dtails afin dillustrer les principales
caractristiques et questions que soulve chacune de ces
couches.
* S.J.D. Candidate and Graduate Fellow, Centre for Innovation Law & Policy, Faculty of Law,
University of Toronto. E-mail: craig.mctaggart@utoronto.ca.
McGill Law Journal 2003
Revue de droit de McGill 2003
To be cited as: (2003) 48 McGill L.J. 571
Mode de rfrence : (2003) 48 R.D. McGill 571
572
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
Introduction
I. A Layered Conceptual Model of Internet Architecture
A. Monolithic vs. Layered Network Architectures
B. Layers and Law
C. A Four-Layer Conceptual Model of Internet Architecture
1. Physical Layer
a. Equipment Sublayer
b. Networks Sublayer
2. Operational Layer
3. Application Layer
4. Content Layer
a. Content Sublayer
b. Transactions Sublayer
a. Centralized Resources and Functions Sublayer
b. Standards and Protocols Sublayer
c.
ISP Functions Sublayer
II. A Layered Typology of Internet Legal and Policy Issues
A. Content Layer
1. Content Sublayer
2. Transactions Sublayer
3. Example: Browse-Wrap Licences
B. Application Layer
1. Example: Overlay Software
C. Operational Layer
1. Centralized Resources and Functions Sublayer
2. Standards and Protocols Sublayer
3.
ISP Functions Sublayer
4. Example: E-Mail Service
D. Physical Layer
1. Equipment Sublayer
2. Networks Sublayer
3. Example: IP Telephony
Conclusion
Glossary
573
574
574
576
580
583
583
583
584
584
586
586
587
587
587
588
588
590
591
592
594
597
600
603
604
606
608
611
615
616
616
619
622
625
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
573
Introduction
The analysis of Internet legal and policy issues is aided by an understanding of
the Internets unique layered architecture. Instead of being viewed as a monolithic
whole, the Internet can be thought of as being composed of four different conceptual
layers; from the top down these are: the content, application, operational, and
physical layers. While Internet content (e.g., World Wide Web pages) and
transactions (e.g., e-commerce) are its most familiar aspects, there is much more to
the Internet than meets the eye. Many different elements within each of its four layers
must work together to make Internet content appear on users computer screens.
Briefly, these layers can be defined as follows:
Content Layer
The data available by means of the Internet and
transactions enabled by the Internet.
Application Layer
Operational Layer
The software applications that make Internet
content available and that enable Internet
transactions.
The centralized resources and functions,
standards and protocols, and Internet Service
Provider (ISP) functions essential to Internet
operations.
Physical Layer
The computer equipment and telecommunications
networks over which the Internet operates.
the Internet
is not
Contrary
to early popular perceptions,
inherently
uncontrollable. Rather, each constituent element is subject to varying patterns of
control by a variety of different parties. Addressing these patterns of control through
Internet legal and policy analysis necessitates a degree of technical precision. This
article proposes a layered conceptual model of the Internet that can be used to place a
vast range of Internet-related legal and policy issues into their appropriate contexts.
Many such issues are canvassed and some are analyzed in detail in order to provide
illustrations of this approach.
The purpose of this article is to help decision-makers, policy-makers, and legal
analysts better comprehend the Internet, to ensure that they identify, with an
appropriate degree of granularity, the precise issue before them and consider its
possible links to, and implications for, other elements of the Internet. The layered
model is put forward as a particular way of thinking about the Internet, not as a way
of solving any particular problem. It is hoped that readers will find this conception
useful in their own study of Internet legal and policy issues.
574
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
This article is premised on the idea that the Internet is important and will continue
to increase in importance as a social phenomenon. In spite of the dot-com
implosion of early 2000 that followed the Internets meteoric rise to ubiquity in the
late 1990s, Canadian Internet-use rates continue to climb.1 The disappointment of e-
commerce appears to have done nothing, for example, to dampen interest in e-mail,
illustrating the importance of distinguishing one Internet application from another. As
was the case with other revolutions in communications, such as the telephone and
television, the aphorism that people tend to overestimate the short-term impact of new
technologies and underestimate their long-term impact may be borne out by the
Internet, or at least by Internet technology.2 If that is the case, then the short-term
impact of the Internet on our legal system may only be the beginning of much greater
challenges.
Part I of this article provides an introduction to the Internets four conceptual
layers and many of their respective sublayers and elements. Part II offers a typology
of representative legal and policy issues organized by the layer and sublayer to which
each issue primarily (though not exclusively) relates. Issues associated with each
sublayer are surveyed, and one example with respect to each layer is discussed in
detail. I observe the considerations that typify each layer as well as explore their
respective policy implications. In the conclusion I contemplate future applications of
the layered approach to Internet legal analysis.
I. A Layered Conceptual Model of Internet Architecture
To appreciate the significance of the Internets layered architecture, one must first
understand how the Internets architecture differs from that of previous public
communications infrastructures.
A. Monolithic vs. Layered Network Architectures
In the classical North American telephone system of the monopoly era, the
phone company controlled virtually all aspects of telecommunications within its
territory. Starting at the bottom, their comprehensive responsibility included digging
the trenches, stringing the cable, and connecting network facilities and subscribers.
Subscribers could only use devices supplied, if not also manufactured, by the phone
company. Phone numbers were assigned by the phone company and listed in the
phone companys directory. For most of the history of the telephone, the user could
only use it for one purposevoice telephone calls. While it may seem trite, this
1 Statistics Canada reports that [m]ore than 5.8 million households, or 49% of all 12 million
households, had at least one member that regularly used the Internet from home in 2001, up 1.1
million (+23%) from 2000. This was somewhat less than the gain of 1.4 million (+42%) from 1999
to 2000 (Statistics Canada, Household Internet Use Survey: 2001 The Daily (25 July 2002),
online: Statistics Canada
2 This maxim is widely credited, without citation, to science fiction writer Arthur C. Clarke.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
575
limitation, and the structural reasons for it, become significant when contrasted with
the open, layered data networks of the Internet.
So long as one company was in complete control of all aspects of the telephone
network, its use was narrowly defined. The network was very good at carrying voice
telephony, whether down the street or around the world. Carriers made physical and
operational layer connections with other carriers. Each carrier, however, also
controlled that portion of the overall telephone system that was physically located
within its territory. So extensive was this control that until 1968 in the United States,
and 1982 in Canada, the telephone companies could prohibit the connection of
equipment that they did not make or approve of to their networks (referred to as
foreign attachments).3 In the monopoly era, public telecommunications networks
could indeed be thought of in monolithic terms.
This situation was
the sea of change
transformed, however, by
in
telecommunications (and its regulation) that followed the replacement of monopoly
principles with market principles in Canadian telecommunications policy.4 With the
competitive provision of enhanced services such as voice mail, data processing,5
and long distance services,6 enterprises unrelated to telephone companies were given
greater access to the latters facilities on regulated terms. In addition, developing
alongside (and thanks to) these changes was the lessening of the telephone
companies control over what services could be provided by means of the system.
This development was key in the rise of public data networking. Monolithic network
control gave way to decentralized control of separate elements at separate layers.
Once almost anything could be connected to the telephone system, a remarkable
era of innovation in telecommunications and information technologies began. The
telephone companies continued to hold monopolies (whether de jure or de facto) over
many elements of the telephone system such as the physical access infrastructure, but
generally they could not restrict the applications using that infrastructure or the
content passing over it. Facsimile machines, speaking in standardized analog tones,
were an early manifestation of this freedom. Private data networks, employing
standardized digital protocols, were another. Large enterprises began to interconnect
3 See Carterphone v. American Tel. & Tel. Co., 13 F.C.C. 2d 420 (U.S. Federal Communications
Commission 1968); Attachment of Subscriber-Provided Equipment (23 November 1982), Telecom
Decision CRTC 82-14. Canadian Radio-television and Telecommunications Commission (CRTC)
documents from approximately 1996 onward are available online: CRTC
4 For a detailed account of this transition, see John S. Tyhurst, Monopoly Lost?: The Legal and
Regulatory Path to Canadian Telecommunications Competition, 1979-2002 (2001/2002) 33 Ottawa
L. Rev. 385.
5 Enhanced Services (12 July 1984), Telecom Decision CRTC 84-18, online: CRTC
6 Competition in the Provision of Public Long Distance Voice Telephone Services and Related
Resale and Sharing Issues (12 June 1992), Telecom Decision CRTC 92-12, online: CRTC
576
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
scattered computing facilities to exchange internal data traffican early step towards
the user taking control of what public telecommunications networks could do.
included
The technical architecture of todays Internet was first developed in U.S.
government-sponsored research networks during the 1970s and 1980s. These
networks
the U.S. Advanced Research Projects Agency Network
(ARPANET), under the aegis of the U.S. Department of Defense and the U.S.
National Science Foundation Network (NSFNET). Access to these networks was
initially restricted to the American defence establishment and the government-funded
scientific community, respectively.7 The rules governing the NSFNET were relaxed in
the early 1990s so that public, commercial networks could interconnect with it.8 This
step was the beginning of the Internets commercial era, in which anyone with an
Internet connection could send e-mail or put up a Web site, among other empowering
benefits.
The fundamental difference between data networks such as the Internet and the
telephone networks that preceded it is the separation of control over the use of the
network from control over the network itself. As will be explained in the next section,
control over the physical, operational, and application layers (i.e., those layers below
the content layer) is further separated within the Internets infrastructure. Control over
individual elements is sometimes even more dispersed. This fragmentation of control
is what lies behind the idea that nobody controls the Internetno one entity is in a
position to control every element of every layer. To put it another way, nobody can
turn it off.9
B. Layers and Law
Much like the Internets potential to change the rules of business, the extent to
which it could exist outside real-world law was overhyped in the 1990s. The title of
Canadas first law journal article on the subject, Controlling the Uncontrollable:
Regulating the Internet by Dov Wisebrod,10 expressed the popular notion at the time
that the Internet either could not be controlled or would prove particularly resistant to
regulation by the state.11 After offering the view that King Canute had as much
7 The fascinating history of these early internetworks can be easily accessed in Katie Hafner &
Matthew Lyon, Where Wizards Stay Up Late: The Origins of the Internet (New York: Simon &
Schuster, 1996).
8 See Neil Randall, The Soul of the Internet: Net Gods, Netizens and the Wiring of the World
(London: International Thomson Computer Press, 1997) at 248-50.
9 Brian Carpenter, ed., Request for Comments (RFC) 1958Architectural Principles of the
Internet (June 1996) at s. 2.4, online: Internet Engineering Task Force
10 Dov Wisebrod, Controlling the Uncontrollable: Regulating the Internet (1995) 4 Media and
Communications Law Review 331.
11 Perhaps the classic academic expression of this idea is found in David R. Johnson & David Post,
Law and Borders: The Rise of Law in Cyberspace (1996) 48 Stan. L. Rev. 1367.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
577
success commanding the tides to retreat as a national government will have regulating
cyberspace,12 Wisebrod argued:
The very essence of the Internet is anarchy, a diametrical opposite of
authority. … [T]he anarchy of the Internet is a powerful, cooperative,
functional force that cannot be subjected to centralized control. Thus, while
the existence of a normative basis for regulating the Internet would be an
interesting subject for debate, the exercise has little practical application. Due
to the nature of the Internet, including its history, culture, amorphousness, and
universality, it is quite impossible to effectively regulate.13
While the Internet may appear anarchical in the context of Usenet,14 which is the
single, text-oriented application that formed the object of Wisebrods analysis,15 the
physical and operational layers are far from anarchicalindeed, they cannot be if the
Internet is to function. The centralized resources and functions must be stable and
non-conflicting. Internet Service Providers (ISPs) must carefully control and
coordinate the delivery of traffic. Eight years of experience since Wisebrods ground-
breaking article have shown that it is simply not possible to make blanket statements
as to whether the Internet is, or can be, controlled and regulated. The answer to that
question depends entirely on the specific aspect of the Internet at issue. More
precisely, it depends on the specific element of the specific layer and sublayer at
issue.
Wisebrod later qualifies the term anarchy to mean co-operative anarchy.16
That the Internet is a co-operative environment is an important insight.17 One must
add to this, however, Lawrence Lessigs insight that, far from being anarchic, the
Internet can be regulated as much by the design of software code as by statute and
indeed is already regulated in many such ways. In Code and Other Laws of
Cyberspace18 Lessig refutes the idea that the Internet has an essence or a fixed
nature. Rather, its shape and function are determined by those in control of its
architecture:
12 Wisebrod, supra note 10 at 332.
13 Ibid. at 332-33.
14 Usenet (short for USEr NETwork) is defined as follows: A public access network on the Internet
that provides user news and group e-mail. It is a giant, dispersed bulletin board that is maintained by
volunteers who provide news and mail feeds to other nodes. Alan Freedman, Computer Desktop
Encyclopedia, 9th ed. (New York: Osborne/McGraw-Hill, 2001), s.v. Usenet.
15 Wisebrod explicitly notes, The remainder of this article assumes the Internet is comprised
primarily of Usenet (supra note 10 at 337).
16 Ibid.
17 While we are beginning to learn more about the nature of the relationships between ISPs and their
customers, there has been very little research on the relationships among ISPs. So mysterious is this
realm that it is often depicted in popular and even technical literature simply as a cloud. Given the
Internets social and economic importance, at least in North America, this obscurity cannot likely
continue much longer.
18 Lawrence Lessig, Code and Other Laws of Cyberspace (New York: Basic Books, 1999) [Lessig,
Code].
578
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
As the world is now, code writers are increasingly lawmakers. They determine
what the defaults of the Internet will be; whether privacy will be protected; the
degree to which anonymity will be allowed; the extent to which access will be
guaranteed. They are the ones who set its nature. Their decisions, now made
in the interstices of how the Net is coded, define what the Net is.
How the code regulates, who the code writers are, and who controls the code
writersthese are questions that any practice of justice must focus [on] in the
age of cyberspace. The answers reveal how cyberspace is regulated.19
In Future of Ideas, Lessig develops this theory and argues that commercialization
is radically changing the Internet, such that the commons it represents is being
enclosed in many ways:
[A]t just the time that the Internet is reminding us about the extraordinary
value of freedom, the Internet is being changed to take that freedom away.
Just as we are beginning to see the power that free resources produce, changes
in the architecture of the Internetboth legal and technicalare sapping the
Internet of this power. Fueled by a bias in favor of control, pushed by those
whose financial interests favor control, our social and political institutions are
ratifying changes to the Internet that will reestablish control and, in turn,
reduce innovation on the Internet and in society generally.20
Lessigs purpose is much broader than mine. I offer no opinion here on whether
control is good or bad, but merely aim to draw attention to the way the Internets
unprecedented architecture gives rise to varied issues of law and policy at each layer.
This article will serve its purpose if it assists those dealing with such issues to better
understand both the context and implications of their decisions.
With a few notable exceptions, early Internet legal literature focussed on one
layer and either played down or completely ignored the rest.21 Most legal writing and
media coverage to date has been related to the content layer. There is, however, much
more to the Internet than just its top layer. Content and transactions are the
finished product, but they could not exist without the lower layers, the
infrastructure.
Canadian Tim Wu was the first scholar to argue for something of an architecture-
influenced approach to Internet legal analysis. In a 1999 article, Wu writes:
For most purposes, I think we ought to discard the old talk of the Internet as a
whole, for the whole Internet is rarely an appropriate level on which to
generalize. Instead, legal thinking can better focus on where the variation that
is apparent to the user is actually found: the application layer above the
19 Ibid. at 60.
20 Lawrence Lessig, The Future of Ideas: The Fate of the Commons in a Connected World (New
York: Random House, 2001) [Lessig, Future of Ideas] at 15.
21 See e.g. I. Trotter Hardy, The Proper Legal Regime for Cyberspace (1994) U. Pitt. L. Rev.
993; James Boyle, Foucault in Cyberspace: Surveillance, Sovereignty, and Hardwired Censors
(1997) 66 U. Cin. L. Rev. 177; Johnson & Post, supra note 11; Wisebrod, supra note 10 at 332-33.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
579
Internets basic protocols. We need, I think, to focus on the user, not on the
network, and that means legal analysis that begins with the application.22
Wu suggests that Internet legal analysis proceed on an application-by-application
basis (e.g., the Web, e-mail, video streaming). He continues:
Whats the difference? This seemingly technical point matters because the
Internet by its design allowseven encouragesgreat diversity above a few
basic standards. The end-to-end design of the Internet delegates the power
to code function to the point nearest to the user: the application. As a result,
nearly everything that counts about the Internet from a legal standpoint is a
function of the particular application at issue and not of the basic Internet
protocols. Since applications actually drive Internet usage, they ought also
drive legal analysis of the Internet, yielding nuanced rather than stereotyped
results.23
While I endorse this argument to a point, note that it obscures at least two
important aspects of the Internet. First, it omits entirely an account of the physical
layer, on which all higher layers ride. Second, Wus argument fails to recognize the
complexity of issues that arise at the operational layer. There is much more to what
makes the Internet work than a few basic standards. Wus view either assumes that
all the elements of the Internet below applications are uninteresting or uncontroversial
or declares that applications, for the purposes of legal analysis, are the Internet. In the
model proposed in this article, the application layer (further divided into client-side
and server-side manifestations) is recognized as a distinct layer for the purposes of
legal analysis; however, it is one of four layers and is not the only layer of import.
A second model was proposed by Professor Yochai Benkler, who employs a
tripartite model comprising a content layer, a logical layer, and a physical
layer.24 While his model goes further to affirm the significance of what I call the
operational and physical layers, it nevertheless generalizes too much about the logical
layer and neglects to identify the distinct sublayers within the logical layer, such as
the centralized resources and functions and the ISP functions. I therefore prefer
operational layer to logical layer as a broader term. Finally, I suggest that
Benklers top layer be subdivided into content and transaction sublayers in order to
embrace both static content and interactive e-commerce transactions.
In Future of Ideas,25 Lawrence Lessig adopts Benklers three-layered model but
refers to Benklers logical layer as the code layer, to pick up on the title of his own
22 Timothy Wu, Application-Centered Internet Analysis (1999) 85 Va. L. Rev. 1163 at 1164 [Wu,
Internet Analysis].
23 Ibid.
24 Yochai Benkler, From Customers to Users: Shifting the Deeper Structures of Regulation Toward
Sustainable Commons and User Access (2000) 52 Fed. Comm. L.J. 561 at 562.
25 Lessig, Future of Ideas, supra note 20 at 23-25.
580
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
book, Code and Other Laws of Cyberspace.26 By implying that code-layer software
makes the hardware run on its own, Lessigs model continues the trend of
obscuring the critical roles played by the Internets centralized resources and
functions and by its ISPs. Lessigs formulation of Internet layers also conflates the
Internets standards and protocols (the protocols that define the Internet) with the
applications that employ them (the software upon which those protocols run).27 In
my view, these two types of code perform such different functions and relate to such
distinct functional aspects of the Internet that they warrant division into separate
layers. As illustrated above by reference to earlier public telecommunications
network paradigms, it is the Internets unique user-definable application layer that
makes it special among communications networks. This feature is what Wu is
referring to when he says that the technical design of the Internet delegates the
power to code function to the point nearest to the user: the application.28 Code that is
user-definable (e.g., application software) is therefore qualitatively different from
code that is standardized (e.g., operational standards and protocols).
Another way to distinguish these two types of code is to think of application-
layer code as edge code and operational-layer code as core code. It is an Internet
industry convention to refer to those parts of the Internet that are closest to users as
the networks edge29 and to everything in between these edges as the core. As
such, edge code consists mainly of application software and is user-definable, while
core code consists mainly of standards and protocols with which the user normally
has almost no interaction. Examples of core code and edge code are provided below
in the sections describing the operational and application layers respectively.
C. A Four-Layer Conceptual Model of Internet Architecture
Bearing in mind the way in which the present model is distinguished from
previous models, the four conceptual layers of the Internet discussed in this article
are, from bottom-up: the physical layer, the operational layer, the application layer,
and the content layer.30 Below is a graphical representation of these four layers, their
26 Lessig, Code, supra note 18. Code is a computer industry term for software or the individual
machine instructions of which it is composed.
27 Lessig, Future of Ideas, supra note 20 at 23.
28 Wu, Internet Analysis, supra note 22. Note that code is used as both a noun and a verb in the
computer industry.
29 Though there are, of course, no edges to the Internet at all, only nodes in a vast mesh.
30 I note that two unpublished papers have been presented at the annual Telecommunications Policy
Research Conference in Washington, D.C. in recent years that apply somewhat similar four-layer
models to U.S. cable and telecommunications regulation. See Kevin Werbach, A Layered Model for
Internet Policy (Paper presented to the 28th Research Conference on Information, Communication,
and Internet Policy, September 2000) [unpublished], online: University of Michigan
Telecommunications Policy (Paper presented to the 30th Research Conference on Information,
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
581
respective sublayers, and some examples of their constituent elements. It is important
to bear in mind that this model is not intended to be comprehensive. Furthermore, this
model is purely conceptualInternet architecture does not conform exactly to these
divisions in reality. The rationale for the model is further explained at the outset of
Part II.
Communication, and Internet Policy, September 2002) [unpublished], online: University of Michigan
582
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
Four-Layer Conceptual Model of Internet Architecture
Layers
Sublayers
Examples of Elements
Content Layer
Content
Transactions
Client-side Applications
Application
Layer
Server-side Applications
Web pages
Digital media, software files
E-mail messages
E-commerce transactions
Interactive transactions
IP telephony, instant messaging,
and e-mail client software
Web browser software
Peer-to-peer file-sharing
applications
Multi-player video games
IP telephony, instant messaging,
and e-mail server software
Web server software
Multi-player video game servers
Centralized Resources
and Functions
Domain names
IP addresses
Root server system
Operational
Layer
Standards and Protocols
TCP/IP, HTTP, SMTP, BGP*
ISP Functions
Equipment
Networks
Physical Layer
Interconnection and routing
Naming and addressing
Mail and name server operation
Computers (terminals
and servers)
Routers
Telephone, cable TV networks
Wireless networks
*
See the glossary at page 625, below, for the meanings of these acronyms.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
583
1. Physical Layer
The physical layer can be defined as: (a) the computer equipment, and; (b) the
telecommunications networks over which the Internet operates.
a. Equipment Sublayer
Internet equipment consists mainly of computers functioning as terminals,
servers, or routers. Terminals are those computers used at the edges by users and
today consist largely of Personal Computers (PCs). Servers are the core computers
that host data and perform operational functions. Routers are computers as well but
with a more specialized function: to deliver packets, the basic unit of Internet traffic
consisting of chopped up messages and data, to other routers.
In the case of terminals, servers, and routers, their physical manifestation is
separate from their operational control as a result of their modular design. User
terminals and provider servers run applications that are under the control of the user
and provider, respectively. Router software is controlled by the ISPs that operate
them. Most of these control functions can be performed remotely over the Internet,
making physical control of the equipment less important than operational control. The
consequences of this modular pattern of control are discussed in the layered typology
set out in Part II.
b. Networks Sublayer
It is often forgotten that the Internet operates atop other telecommunications
networks, including the Public Switched Telephone Network (known in the
telecommunications world as the PSTN). Telephone and Internet traffic are often
carried over the same physical wires and cablesthey are simply encoded using
different operational protocols. In the classical North American Internet access model
of the 1990s, the humble copper loop PSTN line was the predominant type of first
mile physical layer link connecting the user with the Internet. But the copper loop is
just one type of physical link. Coaxial cable networks and wireless radio frequency
networks can also carry data. Wireless local area networks31 are increasingly being
used to provide always-on wireless Internet connectivity. Even electric power
transmission lines are capable of serving as the physical layer for communications
networks.32 Fibre optic cable has now replaced metallic wire in the long-haul portions
31 Based on the Institute of Electrical and Electronics Engineers 802.11 networking standards.
32 See Julia Scheeres, Net Access: Socket to Me Wired News (17 April 2001), online: Wired News
584
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
of most telecommunications networks, and there is a push in some quarters to extend
the reach of high-capacity (and therefore high-speed) fibre closer to users.33
All of this said, it is worth briefly noting some elements below the physical layer
(though implicit in it for present purposes), such as rights-of-way, trenches, and
support structures (e.g., poles and conduits), which are simply assumed to exist in
most discussions of networks, but for which the real-world constraints of space,
climate, and distance should not be forgotten. Distance may be dead at the operational
layer,34 but it is still an important consideration at the physical layer.35
2. Operational Layer
The operational layer, made up of the Internets centralized resources and
functions, standards and protocols (or core code), and ISP functions, is the most
obscure part of the Internet from the users point of view, but it is the essential glue
which binds all the other layers together. The physical, application, and content layers
could exist separately without the operational layer, but they could not be
interconnected without it. There are three major sublayers of operational-layer
elements, and thus areas in which distinct legal and policy issues arise: (a) the
Internets centralized resources and functions; (b) the standards and protocols to
which participating networks adhere; and (c) the operational functions performed by
ISPs of all sizes.
a. Centralized Resources and Functions Sublayer
Domain names, and their numerical cognates, Internet Protocol (IP) addresses,
are used by ISP facilities to route Internet traffic (in packet form) among
interconnecting ISP networks from origin to destination, in response to user
commands. A hierarchical system of servers, known as the Root Server System
(RSS), makes available certain information that binds domain names to IP
addresses. In this way, routers know where to send packets when the user has
requested a Web page with a particular Uniform Resource Locator (URL) or sent an
e-mail to a particular e-mail address. Each ISP assigns IP addresses to its Internet-
connected equipment and manages the assignment of domain names to its customers
servers. The ISP shares this information with other ISPs in the complex
33 See Canadas Advanced Internet Development Organization (CANARIE), Gigabit Internet
to the Home and Schools, online: CANARIE
34 Frances Cairncross, The Death of Distance: How the Communications Revolution Will Change
Our Lives (Boston: Harvard Business School Press, 1997).
35 As illustrated by several protracted right-of-way disputes between carriers and municipalities in
recent years. See e.g. Ledcor/VancouverConstruction, Operation and Maintenance of Transmission
Lines in Vancouver (25 January 2001), Decision CRTC 2001-23, online: CRTC
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
585
interconnection and routing process that makes its customers servers accessible
across the Internet.36
The nature of digital network identifiers (i.e., domain names, numbers, and
addresses) requires that they be globally unique in order to be resolvable or useful
across the network that employs them. It is an unavoidable technical constraint that
there can only be one of each domain name, number, or address on a network. This
constraint means that there can only be one of each domain name on the global
Internet. The resulting clash between trademarks (which can co-exist on Earth in spite
of their similarity) and domain names (which cannot) has resulted in volumes of
scholarly writing, court and arbitral decisions, and regulatory activity (mainly in the
U.S.). The problem, in the minds of many, has yet to be resolved satisfactorily. I
further discuss this problem, along with several other issues related to the centralized
resources and functions, in the layered typology below.
This restrictiveness of network identifiers stands in stark contrast to the bountiful
diversity of the content layer and renews debates over scarcity that many thought the
Internet would render irrelevant. Andeen and King explain the reality as follows:
Ultimately, the fundamental technical driver of addressing is that Top
Level Domains of any addressing scheme must be under the authority of a
single, superordinate power if the network is to be globally effective. There is
no way to avoid this.37
The Internet Domain Name System (DNS) is a hierarchical system of names and
databases, and as the quotation above suggests, someone has to hold ultimate
authority over the top level of this hierarchy.38 For most of the Internets history, this
power was held by the U.S. government but exercised by a small team of engineers at
the University of Southern California under the leadership of computer scientist
Jonathan Postel.39 His role was institutionalized in 1998 with the Internet Corporation
for Assigned Names and Numbers (ICANN).40
36 More is said about the roles played by ISPs after an explanation of Internet standards and
protocols in Part I.C.2.c, below.
37 Ashley Andeen & John Leslie King, Addressing and the Future of Communications
Competition: Lessons from Telephony and the Internet in Brian Kahin & James H. Keller, eds.,
Coordinating the Internet (Cambridge: Massachusetts Institute of Technology Press, 1997) 208 at 251.
38 Despite a preponderance of expert technical opinion in favour of a single Internet DNS hierarchy
(see e.g. Internet Architecture Board, RFC 2826: IAB Technical Comment on the Unique DNS
Root
Internet Engineering Task Force
circumstances should, coexist. See Milton L. Mueller, Competing DNS Roots: Creative Destruction
or Just Plain Destruction? (Paper presented to the 29th Research Conference on Information,
Communication, and Internet Policy, 28 October 2001) [unpublished], online: Cornell University
(May 2000), online:
39 As a 25-year-old graduate student at the University of California at Los Angeles, Jon Postel
volunteered to maintain the authoritative lists of many parameters in use on the ARPANET, which
eventually came to include domain names. Postel held various roles (under the semi-official title of the
586
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
b. Standards and Protocols Sublayer
Transmission Control Protocol/Internet Protocol (TCP/IPoften abbreviated to
simply IP) is the language or protocol that all computers on the Internet speak.
TCP/IP is the most important element of the Internets core code, though it is not the
only one. TCP/IP is an end-to-end protocol suite, meaning that any Internet-
connected computer is capable of speaking directly with any other connected
computer and by means of any one of a number of different protocols. By contrast,
telephones exist in a master/slave relationship with the switch to which they are
connected. Internet packets need to be switched too, but a dedicated circuit between
sender and receiver is not required: the packets merely join the rivers of other packets
traversing the multitude of interconnected networks that comprise the Internet.
TCP/IP makes sure that, eventually, they all get to where they are supposed to go.41
The TCP/IP suite consists of three general protocols (TCP, IP, and User Datagram
Protocol) in addition to many others with specific functions.42 Each protocol provides
a simple, standardized way to participate in the global Internet. The protocols are
voluntarythat is, no one requires any Internet-connected network to use any
particular protocolbut the price of non-compliance is isolation. In another stark
contrast to the diverse content layer, standards and protocols require absolute
conformity. The question of who sets these powerful standards and on what basis is
one of the most important public policy questions relating to the Internet.
c.
ISP Functions Sublayer
Simply having a physical link to a telephone or cable TV network is not enough
to connect to the Internet. One needs to enter into a relationship with a specialized
intermediarycommonly known as an Internet Service Provider (ISP). In the
classical North American dial-up Internet access model, the users modem dials a
telephone number that an ISP has designated to receive incoming calls to its own
modems. The devices shake hands and speak to each other in order to establish a
two-way TCP/IP connection over which packets can pass. With broadband
connections, slightly different hardware performs the same functions. ISPs also
provide other services such as the operation of domain name resolution software and
the hosting of e-mail, Web, and application servers. Some provide extra security
1998),
online:
Force
Internet Assigned Numbers Authority (IANA)) under U.S. government contracts until his untimely
death on 16 October 1998. See V. Cerf, Request for Comments (RFC) 2468: I Remember IANA (17
October
Internet Engineering Task
40 The ICANN Web site is found at
41 A more detailed explanation of TCP/IP and its origins is found in Randall, supra note 8, c. 4,
Vint Knows Protocol: The Birth of TCP/IP. See also Freedman, supra note 14, s.v. TCP/IP.
42 Examples listed in the table at page 582, above, include routing protocols such as Border
Gateway Protocol (BGP) and application-supporting protocols such as Simple Mail Transfer Protocol
(SMTP) and Hypertext Transfer Protocol (HTTP).
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
587
services such as spam filtering, a server-side measure intended to keep junk e-mail
out of customers inboxes.
3. Application Layer
The application layer consists of the software applications that make Internet
content available and that enable Internet transactions. Such software allows users to
post, send, and receive information of all kinds via the Internet. These applications
underlie the content layer and rely on the availability of the operational layer below.
The malleability of edge code allows any two (or more) users to communicate over
the Internet by means of any software application they wish, be it commercial or
homemade. Instead of just taking part in telephone calls, as with the classical
telephone network, users can define what the network does for them. As suggested
above, this development is truly revolutionary in public communications networking.
Internet applications have client-side and server-side
manifestations;43 that is, it takes software on both a user terminal and a provider
server to effect a two-way exchange of data.44 Common examples of application layer
software include Web browser and server software, multi-player video games, and
client and server software for IP telephony, instant messaging, and e-mail. There are
literally hundreds of other applications that support such Internet-based functions as
software upgrades, video conferencing, and remote equipment monitoring.
Almost all
4. Content Layer
The content layer is the layer the user sees: it is the reason he or she uses the
Internet in the first place. It comprises the data available by means of, and
transactions enabled by, the Internet.
a. Content Sublayer
Technically speaking, all Internet content exists in the form of computer data
stored on servers until it is called up by a client-side application and arranged for
display on a users screen. The more common meaning of content, and the one used
in this article, is simply the information that the user can access via the Internet. The
range of such information is, of course, incomprehensibly vast. Internet content can
be anything digitizable: any information that is convertible into digital form, or, as
with most modern media, information that is created in digital form. The most
43 Client/server is [a]n architecture in which the users PC (the client) is the requesting machine
and the server is the supplying machine, both of which are connected via a Local Area Network
(LAN) or Wide Area Network (WAN). Freedman, supra note 14, s.v. client/server.
applications, introduced in Part II.B, below.
44 The fascinating exception
is second-generation server-less peer-to-peer file-sharing
588
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
common forms of content include Web sites, audiovisual media, software, and e-mail
messages.
b. Transactions Sublayer
Internet transactions include electronic interactions of all kindswhether
commercial, non-commercial, human, or machineover the Internet (e.g., gambling,
buying and selling goods and services, submitting a job application). Internet
transactions can be instantaneous or they can result, for example, in the delivery of a
good, either in digitized form over the Internet or in physical form by some real-
world means of delivery. Following a typological exposition of Internet issues, I will
return to transaction issues and specifically to the example of browse-wrap
licences.
II. A Layered Typology of Internet Legal and Policy Issues
This section offers a typology of representative Internet legal and policy issues.
The issues are organized according to the conceptual layers and sublayers to which
they primarily (though not exclusively) relate. In order to create this typology, I
survey a number of issues associated with each sublayer, and one example per layer is
discussed in depth. Some of the considerations that typify each layer are noted. The
descriptions and discussions of each layer and sublayer are intended to place well-
known issues in context and to illuminate some of the many lesser-known issues that
may require attention in the future.
intellectual property
There are three caveats. First, this typology is suggested as one way of
conceptually organizing Internet legal and policy issues; it is certainly not the only
way. One could also classify these issues by dominant legal regime (e.g., commercial
law,
issue area (e.g., digital media, gambling,
telecommunications), or even by jurisdiction. The classification of issues by layer,
however, is intended to demonstrate the degree to which the architectural and legal
circumstances of each layer can differ and to show how control of elements at one
layer can have an impact on the elements at other layers.
law),
Second, in an article of this length it is not feasible to attempt to identify every
aspect of every issue. Indeed, each sublayer gives rise to such a wide range of issues
that the relevant sections below can only refer to a few. This approach is not intended
to downplay the importance of issues that are not mentioned. I have endeavoured to
provide references to further material on those Internet-specific topics that are raised
but that cannot be addressed in detail. In some cases, there have not been any legal or
regulatory proceedings to date, a state of affairs that highlights potential public policy
problems. The Internet will likely continue to give rise to interesting policy issues that
we cannot yet imagine, precisely because we cannot know how people will use the
Internet in the future. This model therefore may be useful as a way of placing such
future questions in context, but does not claim to provide answers to them.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
589
Third, the choice of which layer to advert to when considering a particular issue
is not meant to imply that it does not engage other layersquite the opposite. The
examples discussed in the sections below illustrate the remarkable degree to which
issues cut across layers: they can implicate many different layers and legal regimes in
different ways. Indeed, while it may be possible to make generalizations about some
legal regimes, for example, that copyright law is primarily relevant to the upper two
layers (content and application), and telecommunications law relates to the bottom
two layers (operational and physical), some regimes are so far-reaching that they have
an impact on almost all of the layers.
transactionshealth
information disclosed
Privacy issues, for example, are implicated at each sublayer: contentthe text of
e-mail messages;
in an on-line
pharmaceutical purchase; client-side applicationscookies,45 which track Web
browsing patterns; server-side applicationsWebmail46 e-mail account information
on provider servers; centralized resources and functionsones identity as expressed
in a domain name; standards and protocolsthe presence or absence of privacy-
enhancing features in Internet core code; ISP functionscustomer information such
as logon and logoff times; equipmentunique identifiers on computer hardware;47
and networkswireless network eavesdropping.
Jurisdiction is constantly relevant as an analytical consideration. The Internets
challenge to territorial sovereignty is well known,48 but the examples below illustrate
how jurisdictional patterns can vary not only across layers but even within sublayers.
As Michael Geist has noted, Lurking behind virtually every Internet law issue is the
question of jurisdiction … 49 As such, jurisdiction is not examined on its own here50
but is adverted to in the context of several issues below.
Beginning with the content layer and proceeding back down through the
application, operational, and physical layers, I discuss four specific examples as
follows: content layerbrowse-wrap licences; application layeroverlay software;
45 A cookie is [d]ata created by a Web server that is stored on a users computer. It provides a
way for the Web site to keep track of a users patterns and preferences and, with the cooperation of the
Web browser, to store them on the users own hard disk. Freedman, supra note 14, s.v. cookie.
46 Webmail refers to Web-based e-mail services such as canada.com and hotmail.com where the
users messages are stored on the Webmail providers servers and not on the ISPs servers as with
ordinary Internet e-mail.
47 For example, all ethernet Network Interface Cards (NICs) (the most common means by which
PCs are connected to networks) carry a unique serial number that prevents any two NICs from bearing
the exact same network address.
48 This topic is influentially explored in Johnson & Post, supra note 11.
49 Michael A. Geist, iCraveTV and the New Rules of Internet Broadcasting (2000) 23 U. Ark.
Little Rock L. Rev. 223 at 239 [Geist, iCraveTV].
50 See Michael A. Geist, Is There a There There?: Toward Greater Certainty for Internet
Jurisdiction (2001) 16 Berkeley Tech. L.J. 1345; Vaughan Black & Mike Deturbide, Braintech, Inc.
v. Kostiuk: Adjudicatory Jurisdiction for Internet Torts, Case Comment (2000) 33 Can. Bus. L.J. 427;
Chris Gosnell, Hate Speech on the Internet: A Question of Context (1998) 23 Queens L.J. 369.
590
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
operational layere-mail service; and physical layerIP telephony. References are
made primarily to Canadian materials where possible, and to those in the English-
language common law tradition in particular.51
A. Content Layer
It is appropriate to begin with the content layer, since it is the aspect of the
Internet that is the most familiar to users. The range of Internet content-related legal
issues is limited only by the content that people are willing to put on the World Wide
Web. As in other Internet policy areas,52 the Canadian government led the world in
initiating a comprehensive study of legal issues of liability for content circulating on
the Internet in 1996.53 The excellent discussion of content-related issues in that study
has since been supplemented by a rich array of Canadian legal scholarship on, for
example, hate speech,54 child pornography,55 freedom of expression,56 defamation,57
and broadcasting.58
51 The most comprehensive general resources are: Michael A. Geist, Internet Law in Canada, 3d ed.
(Concord, Ont.: Captus Press, 2002) (casebook) [Geist, Internet Law in Canada]; Barry B. Sookman,
Computer, Internet and Electronic Commerce Law, looseleaf (Toronto: Carswell, 2002) [Sookman,
Electronic Commerce Law]; and George S. Takach, Computer Law (Toronto: Irwin Law, 1998)
[Takach, Computer Law].
52 Such as clarifying the regulatory status of IP telephony (see Part II.D.3, below).
53 Michel Racicot et al., The Cyberspace is not a No Law Land: A Study of the Issues of
Liability for Content Circulating on the Internet (Paper prepared for Industry Canada, February
1997) at 1, online: Industry Canada
54 Francine Aumueller, Hate Propaganda Law and Internet-Based Hate (2001) 44 Crim. L.Q. 92;
Gosnell, supra note 50. Note that the Anti-terrorism Act (S.C. 2001, c. 41, s. 88) amended the hate
messages provision of the Canadian Human Rights Act (R.S.C. 1985, c. H-6) to specifically include
Internet-based speech, a result that had already been achieved somewhat awkwardly under the old
provision in Citron v. Zndel ((2002), 41 C.H.R.R. D/274 (CHRT)).
55 Sanjeev Anand, A Case for Upholding the Child Pornography Law (1999) 25 C.R. (5th) 312;
Tanya Scharbach, Child Pornography in Cyberspace (1996) 2 Appeal 58.
56 Robert Dawkins, Online Liberty: Freedom of Expression in the Information Age (2001) 10 Dal.
J. Leg. Stud. 102; Frank Iacobucci, Recent Developments Concerning Freedom of Speech and
Privacy in the Context of Global Communications Technology (1999) 48 U.N.B.L.J. 189; Gordon
Scott Campbell, Emerging Issues of the Internet and Canadian Criminal Law (1998) 3 Can. Crim.
L. Rev. 101.
57 Randy A. Pepper, Internet Defamation: Canadian vs. American Perspectives (2002) 25
Advocates Q. 190; Jonathon T. Feasby, Who Was That Masked Man?: Online Defamation, Freedom
of Expression, and the Right to Speak Anonymously (2002) 1:1 C.J.L.T., online: Dalhousie
University Electronic Text Centre
Defamation Consciousness (2002) 1:2 C.J.L.T., online: Dalhousie University Electronic Text Centre
Defamatory Sense (2000) 15:1 C.J.L.S. 81; Black & Deturbide, supra note 50; George S. Takach,
Internet Law: Dynamics, Themes and Skill Sets (1999) 32 Can. Bus. L.J. 1 at 19ff.; Craig Martin,
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
591
For the most part, the kinds of legal and policy issues that have arisen in the
Internet content and transaction realm to date resemble old wine in new bottles. When
social phenomena such as home taping of recorded music and gambling went on-
line, following a period of adjustment, the off-line regimes of copyright and criminal
law generally caught up. Indeed, in the area of electronic signatures, it may be the
marketplace that now lags behind the law, as e-commerce has failed to become as
pervasive as many had predicted.
The incredible diversity of content and transactions on the Internet can be
expected to produce diverse legal and policy responses. This diversity can be
contrasted with the operational and physical layers, where uniformity and scarcity are
the hallmarks. At these levels, public law regimes such as telecommunications
regulation and the new Internet governance laws are more prominent. Most users are
only familiar with the Internets content layer. The layered approach suggested here is
designed to highlight the aspects of the on-line network and its related legal and
policy issues that lie below the surface.
1. Content Sublayer
By far the most commercially significant legal regime at the content sublayer is
copyright law, perhaps because the Internet may be the ultimate information
distribution (read: copying) machine. While many fascinating copyright issues are
presented at the content sublayer, my purposes are more circumscribed, so I refer
interested readers to some of the many general works,59 articles,60 and policy
proceedings61 relating to copyright and Internet content in Canada.
Tolofson and Flames in Cyberspace: The Changing Landscape of Multistate Defamation (1997) 31
U.B.C.L. Rev. 127.
58 Jonathan A. Blakey & Justine Whitehead, Approaching a Regulatory Crossroad: Internet
Retransmission Activities in Canada (2002/2003) 3 Internet and E-Commerce Law in Canada 41;
Howard P. Knopf, Internet Copyright Reform Initiatives from Canadian Government (2000/2001) 2
Internet and E-Commerce Law in Canada 33; Geist, iCraveTV, supra note 49. See also Canadian
Heritage (Copyright Policy Branch) & Industry Canada (Intellectual Property Policy Directorate),
Consultation Paper on the Application of the Copyright Acts Compulsory Retransmission Licence
to the Internet (22 June 2001), online: Industry Canada
59 Sunny Handa, Copyright Law in Canada (Markham, Ont.: Butterworths, 2002) [Handa,
Copyright Law in Canada]; John S. McKeown, Fox Canadian Law of Copyright and Industrial
Designs, 3d ed. (Scarborough, Ont.: Carswell, 2000); David Vaver, Copyright Law (Toronto: Irwin
Law, 2000).
60 See e.g. Jeremy F. deBeer, Canadian Copyright Law in Cyberspace: An Examination of the
Copyright Act in the Context of the Internet (2000) 63 Sask. L. Rev. 503; Lisa Anne Katz Jones, Is
Viewing a Web Page Copyright Infringement? (1998) 4 Appeal 60; C. Paul Spurgeon, Digital
Networks and Copyright. Licensing and Accounting for Use: The Role of Copyright Collectives
Evolution or Revolution? (1998) 12 I.P.J. 225; Barry B. Sookman, Copyright and the Information
Superhighway: Some Issues to Think About (1997) 11 I.P.J. 123 & 265; Donald M. Cameron, Tom
592
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
2. Transactions Sublayer
While first-generation e-commerce did not differ significantly from tele-sales
or even catalogue sales before that, the spectre of large volumes of consumer and
commercial transactions moving to the Web has sparked a great deal of interest in the
rules governing on-line transactions. As the Web has matured, more powerful
transactional environments have been enabled, such as auction sites (e.g., ebay.ca),
sophisticated retail storefronts (e.g., chapters.indigo.ca), and government service
windows (e.g., the one that facilitates e-filing of income tax returns in Canada:
netfile.gc.ca). A wide variety of transactions can now be executed over the Internet;
that is, by means of a combination of client-side and server-side application software.
These Web-based systems can inexpensively substitute for specialized, proprietary
networks operated by Electronic Data Interchange (EDI) firms over dedicated
telecommunications networks.62
Today, ordinary browser software and an Internet connection can be used in
conjunction with specialized, but relatively inexpensive, server software to enable
electronic transactions protected by highly secure encryption tools. Even the Society
for Worldwide Interbank Financial Telecommunication (SWIFT), the system used
internationally by major financial institutions for instantaneous fund transfers, is
moving towards an Internet model.63 SWIFT, however, is a private organization, and
the repeated interactions among its members are governed by sophisticated
commercial contracts. This is a very different situation from one-off consumer
contracting over the open Internet, which will be discussed below in the context of
browse-wrap licences.
E-commerce law is that branch of consumer and commercial law relating to
transactions enabled by the Internet and other electronic networks. E-commerce legal
issues64 might be grouped into three areas: (1) the requirements for electronic
documents and on-line contracts,65 (2) consumer protection,66 and (3) the protection
S. Onyshko & W. David Castell, IP on the I-Way (1997) 13 C.I.P.R. 311; Mark B. Eisen, Copyright
and the World Wide Web (1996) 12 C.I.P.R. 405.
61 Industry Canada (Intellectual Property Policy Directorate) & Canadian Heritage (Copyright
Policy Branch), Consultation Paper on Digital Copyright Issues (22 June 2001), online: Industry
Canada
[Industry Canada & Canadian Heritage, Digital Copyright]. See also Knopf, supra note 58.
62 See Brian D. Grayton, Canadian Legal Issues Arising from Electronic Data Interchange (1993)
27 U.B.C. L. Rev. 257.
63 Albeit via a secure virtual private network. See SWIFT, SWIFTs Secure IP Network (SIPN),
online: SWIFT
64 An introduction to the full range of issues and regimes engaged by e-commerce can be found in
Barry B. Sookman, Electronic Commerce, Internet and the Law: A Survey of the Legal Issues
(1999) 48 U.N.B.L.J. 119.
65 John D. Gregory, Canadian Electronic Commerce Legislation (2002) 17 B.F.L.R. 277; Richard
Weiland, The Uniform Electronic Commerce Act: Removing Barriers to Expanding E-Commerce
(2001) 7 Appeal 6; John D. Gregory, Receiving Electronic Messages: Eastern Power v. Azienda
Communale & Ambiente (2000) 15 B.F.L.R. 473; Mark J. Selick, E-Contract Issues and
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
593
of individual privacy.67 Certain types of on-line transactions are subject to existing
regulatory regimes. These transactions include gambling,68 sales of pharmaceutical
products,69 securities transactions,70 and the use of electronic money.71 The
Competition Act72 governs on-line advertising, promotions, contests, and
arrangements among competitors, just as it does off-line.73 The realm of taxation
constitutes yet another major field of transaction-related issues that cannot be
explored in depth here.74
Opportunities for the Commercial Lawyer (2001) 16 B.F.L.R. 1; Ian R. Kerr, Spirits in the Material
World: Intelligent Agents as Intermediaries in Electronic Commerce (1999) 22:2 Dal. L.J. 190; Amy-
Lynne Williams, Electronic Commerce: Cutting Edge Changes and Challenges to Commercial Law
and Practice (1999) 48 U.N.B.L.J. 217; Michael Erdle, Legal Issues in Electronic Commerce
(1996) 12 C.I.P.R. 251; Alfred A. Macchione, Overview of the Law of Commercial Transactions and
Information Exchanges in Cyberspace: Canadian Common Law and Civil Law Perspectives (1996)
13 C.I.P.R. 129.
66 Roger Tass & Kathleen Lemieux, Consumer Protection Rights in Canada in the Context of
Electronic Commerce, online: Industry Canada
Tass & Maxime Faille, Online Consumer Protection in Canada: The Problem of Regulatory
Jurisdiction (2000/2001) 2 Internet and E-Commerce Law in Canada 41; Bradley J. Freedman,
Canadian Provincial Internet Consumer Protection Laws (2001/2002) 2 Internet and E-Commerce
Law in Canada 6; John D. Gregory, Solving Legal Issues in Electronic Commerce (1999) 32 Can.
Bus. L.J. 84; David Waite, Consumer Protection Issues in Internet Commerce (1999) 32 Can. Bus.
L.J. 132.
67 Christopher Berzins, Protecting Personal Information in Canadas Private Sector: The Price of
Consensus Building (2002) 27 Queens L.J. 609; Michael A. Geist, When Dot-Coms Die: The E-
commerce Challenge to Canadas Bankruptcy Law (2002) 37 Can. Bus. L.J. 34; Piero Iannuzzi,
Protecting Transborder Data Flows: A Privacy Model for the 21st Century (2001) 18 C.I.P.R. 337;
John MacDonnell, Exporting Trust: Does E-Commerce need a Canadian Privacy Seal of Approval?
(2001) 39 Alta. L. Rev. 346. See also the resources cited infra note 155.
68 C. Ian Kyer & Danielle Hough, Is Internet Gaming Legal in Canada: A Look at Star Net (2002)
1:1 C.J.L.T., online: Dalhousie University Electronic Text Centre
Internet Gambling and the Canadian Conundrum (2000) 6 Appeal 6.
69 Andrea Faith Russell, Internet Pharmacy: Options for Canadian Regulation (2001) 21 Health L.
70 Anita I. Anand, Securities Law in the Internet Age: Is Regulating by Analogy the Right
Approach? (2001) 27 Queens L.J. 129; Stphane Rousseau, Internet-Based Securities Offerings by
Small and Medium-Sized Enterprises: Attractions and Challenges (2001) 35 Can. Bus. L.J. 226;
Anita I. Anand, A Comment on Internet-Based Securities Offerings by Small and Medium-Sized
Enterprises: Attractions and Challenges (2001) 35 Can. Bus. L.J. 274; Gavin Sinclair, Internet
Direct Public Offerings: New Opportunities for Small Business Capital Finance (2000) 27 Man. L.J.
297.
71 Muharem Kianieff, Show Me the Money!: A Critical Evaluation of Laissez-Faire Internet
Can. 90.
Currencies (2002) 17 B.F.L.R. 215.
72 R.S.C. 1985, c. C-34.
73 Peter Franklyn & Kevin Ackhurst, Competition Law Issues in the New Economy: The
Emergence of B2B Marketplaces (2000/2001) 2 Internet and E-Commerce Law in Canada 49.
74 See Arthur J. Cockfield, Canadas GST E-Commerce Policy (Or How to Catch the Big Fish)
(2002/2003) 3 Internet and E-Commerce Law in Canada 1; Jinyan Li, Rethinking Canadas Source
Rules in the Age of Electronic Commerce (1999) 47 Can. Tax J. 1077 & 1411; Pierre J. Bourgeois &
594
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
3. Example: Browse-Wrap Licences
Can merely visiting a Web site place the user in a contractual relationship with
the Web site publisher? Many of the latter (or at least their solicitors) seem to think
so; they often post detailed licence agreements on their sites purporting to bind
visitors to any number of terms of use. A hyperlink to the terms of these agreements is
often provided at the bottom of Web pages, taking the user to a separate page where
presumably few people other than lawyers or law students ever go. For example, on
the first page of Bell Canadas Ontario region English language residential site, there
are hyperlinks at the bottom labelled Legal, Privacy, and Terms.75 Clicking on
the first hyperlink brings up a ten-section document in the form of a contract. The
first section reads as follows:
1. WEB SITE TERMS AND CONDITIONS. The materials on BELL
CANADAs Website (the Site), which may include text, images, audio
clips, video clips, software and other materials (the Content), are
provided by BELL CANADA for informational purposes only. By
accessing the Site or downloading any Content, you agree to be bound by
the terms and conditions set out below (Terms and Conditions). If you
do not agree to these terms and conditions, do not access the Site or
download any Content.76
The remainder of the terms and conditions are representative of this common type of
Web-based document: disclaimers of liability relating to the sites content and use; a
disclaimer of responsibility for, or endorsement of, non-Bell sites to which links are
provided; restrictions on the use of the site; copyright and trademark notices;
incorporation of the privacy policy; an indemnity in favour of Bell in case of the
users violation of the terms and conditions; notice that the site originates in Canada;
reservation of the right to make changes to the site; and miscellaneous provisions,
including the somewhat presumptuous assertion that [t]he parties have required that
these Terms and Conditions and all related documents be drawn up in English.
These documents generally have two purposes: to put users on notice that certain
uses of a site and its contents are not permitted, and to attempt to shield the Web site
publisher from liability relating to that use. The documents go by several common
names, including terms of use, terms of service, disclaimers, Web-wrap licences, and
browse-wrap licences. Recent judicial decisions in Ontario and the U.S. provide
guidance on the enforceability of these browse-wrap licences.
The browse-wrap concept has its origin in the software industrys shrink-wrap
licence. Except for custom software, commercial software is not sold, but licenced.
Retail copies of commodity software products are licenced on standardized terms.
Luc Blanchette, Income_taxes.ca.com: The Internet, Electronic Commerce, and TaxesSome
Reflections (1997) 45 Can. Tax J. 1127 & 1378.
75 Online: Bell Canada
76 Bell Canada, Customer Care: Legal Notice, online: Bell Canada
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
595
George Takach expresses the prevailing view on the enforceability of shrink-wrap
licences:
[S]ome software companies indicate on the outside of the box that a full
licence agreement is contained within the box. This sort of notice given to the
purchaser prior to or at the time of sale should be sufficient to legally bind the
purchaser to the terms of the full licence inside the box. This is particularly
true given that today these licences have become so standard and ubiquitous
that the average user would be hard-pressed to argue that he did not know
about them or the various terms contained within them.77
Shrink wrap licences are most likely enforceable in Canada, subject to the same
caveat given by the court in the leading American case, ProCD, Inc. v. Zeidenberg:
Shrinkwrap licenses are enforceable unless their terms are objectionable on grounds
applicable to contracts in general (for example, if they violate a rule of positive law,
or if they are unconscionable).78
Today it is just as common for retail software to be delivered in digital form over
the Internet, in which case there is no box or envelope sealed with a sticker that says:
Your use of this software is subject to a licence agreement. Instead, a dialog box is
usually presented to the user during installation asking the user to click on an I
agree icon to signify his or her acceptance of the applicable licence terms. This
evolution became known as the click-wrap licence, and Justice Winkler of the
Ontario Superior Court of Justice confirmed its enforceability in strong terms in
Rudder v. Microsoft.79
Any doubts as to whether a mouse click can satisfy the technical requirements for
contract formation have been removed by the enactment in most Canadian
jurisdictions of statutes based on the Uniform Law Conference of Canadas Uniform
Electronic Commerce Act.80 Subsection 19(1) of the Ontario version, for example,
provides (in part) that:
An offer, the acceptance of an offer or any other matter that is material to the
formation or operation of a contract may be expressed (b) by an act that is
intended to result in electronic communication, such as, (i) touching or
clicking on an appropriate icon or other place on a computer screen 81
But what about browse-wrap licences, such as Bell Canadas, which states: By
accessing the Site or downloading any Content, you agree to be bound by the terms
77 Takach, Computer Law, supra note 51 at 283.
78 86 F.3d 1447 at 1449 (7th Cir. 1996).
79 (1999), 2 C.P.R. (4th) 474, 40 C.P.C. (4th) 394 (Ont. S.C.J.) [Rudder]. In this decision, Winkler J.
ruled that a mandatory arbitration clause contained in a Member Agreement was binding on
representative plaintiffs in a prospective class proceeding despite the fact that they had not read that
portion of the agreement.
80 (1999), s. 20(1)(b), online: Uniform Law Conference of Canada
81 Electronic Commerce Act, S.O. 2000, c. 17, s. 19(1).
596
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
and conditions set out below? This kind of unilateral imposition of licence terms is
probably not enforceable. A recent American appellate decision provides the most
current statement of U.S. browse-wrap law. Sotomayor J. set out the facts and
reasoning in Specht v. Netscape Communications82 as follows:
In order to resolve the central question of arbitrability presented here, we must
address issues of contract formation in cyberspace. Principally, we are asked
to determine whether plaintiffs-appellees (plaintiffs), by acting upon
defendants invitation to download free software made available on
defendants webpage, agreed to be bound by the softwares license terms
(which included the arbitration clause at issue), even though plaintiffs could
not have learned of the existence of those terms unless, prior to executing the
download, they had scrolled down the webpage to a screen located below the
download button. We agree with the district court that a reasonably prudent
Internet user in circumstances such as these would not have known or learned
of the existence of the license terms before responding to defendants
invitation to download the free software, and that defendants therefore did not
provide reasonable notice of the license terms. In consequence, plaintiffs bare
act of downloading the software did not unambiguously manifest assent to the
arbitration provision contained in the license terms.83
While this lesson has surely been learned in the case of sites offering downloads of
software subject to licences (which normally impose a requirement to click on an I
Agree button or the like), the reasoning likely also applies to the browse-wrap
situation, where one can use a Web site without having notice or unequivocally
assenting to any governing terms. Barry Sookmans view, which is supported by the
Specht decision, is:
It is unlikely that a web-wrap agreement will be enforced under Canadian law
unless the user takes an action that manifests assent to the terms of the
agreement or unless the web site operator has made a reasonable attempt to
bring the terms and conditions posted at the site to the attention of the user and
the user has had a reasonable opportunity to read such terms.84
The Ontario Superior Court of Justice decision in Kanitz v. Rogers Cable Inc.85
appears at first glance to run counter to this line of thinking, but on closer inspection
there is a key distinguishing factor. At issue in the case was an arbitration clause
dispute between an on-line service provider and unhappy customers seeking to begin
a class proceeding (as in Rudder). Nordheimer J. upheld the validity of unilateral
amendments to an Internet service agreement, amendments which the plaintiffs
claimed they were not notified about and that removed their ability to seek redress in
82 306 F.3d 17 (2d Cir. 2002).
83 Ibid. at 20.
84 Sookman, Electronic Commerce Law, supra note 51 at 10-18.1 to 10-18.2.
85 (2002), 58 O.R. (3d) 299, 21 B.L.R. (3d) 104 (S.C.J.) [Kanitz]. See also Bradley J. Freedman,
Website Notices of Contract Amendments: Kanitz v. Rogers Cable Inc. (2002/2003) 3 Internet and
E-Commerce Law in Canada 17.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
597
the courts. The service provider posted notice of the changes on its Web site but did
not directly contact the plaintiffs to notify them.
The distinguishing factor in Kanitz, of course, is that there was an existing
contractual relationshipthe users had signed a paper contract before their service
was installed. That contract allowed the service provider to amend the agreement
from time to time, notice of which could be given by posting notice on the
Rogers@Home Web site. The contract urged users to check the Web site periodically
to obtain the latest version of the agreement. A notice indicating that amendments had
been made (but not the nature of the amendments) was found to have been displayed
on the service providers customer support site for a reasonable period of time.
While the Kanitz decision at least signals the Ontario courts willingness to recognize
and enforce Web-based contractual communication, it does not go so far as to make
ordinary browse-wrap terms any more enforceable.
The click-wrap and browse-wrap cases suggest that if Web site publishers want to
subject their visitors to any kind of licence agreement, that agreement must be brought
to the visitors attention. Furthermore, visitors must be provided with a way to signify
their unequivocal assent before gaining access to the material to which the licence
applies. This step would normally be taken when the user wants to interact with the
site beyond mere browsing (e.g., by downloading content or querying a database).
B. Application Layer
The application layer is the realm of edge code, the limitlessly diverse set of
programmes that define what the Internet does for its users. Edge code may be
contrasted with core code, the rigidly uniform standards and protocols facilitating the
exchange of information between and among applications. Edge code can be
conceptually divided further into client-side and server-side applications. Generally,
both a client-side and server-side application are necessary to complete an Internet
communication (e.g., sending an e-mail or calling up a Web page). Nevertheless, each
type of application plays a very different role, and their control is similarly separated.
Client-side software is normally under the control of users, who can configure it
according to their liking. Server-side software is under the control of service providers
or Web site publishers. Peer-to-peer file-sharing applications present a challenging
exception to this paradigm, as noted below, by essentially making every client a
potential server.
Internet content and transactions have attracted the most attention to date, but it is
useful to consider the role that the supporting applications play in some of these
disputes. In this section, after brief note is made of encryption tools, patentability of
e-commerce systems, and peer-to-peer file-sharing applications, the specific case of
overlay software is examined in more detail. While these types of issues are
598
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
commonly analyzed as content matters, thinking about them at the application layer
and in the light of the Internets architecture can add a useful new perspective.
On the topic of client-side application software, the case of encryption is
particularly illuminating. Given the inherent insecurity of communications across the
open Internet86 (as opposed to, for example, the closed EDI environments referred to
above), users must take special measures if they want to protect the privacy and
security of their information exchanges. Encryption technology (to oversimplify
greatly) scrambles and selectively unscrambles data and allows as few as two users to
access a particular communication (be it an e-mail message or a Web banking
deposit). Encryption can therefore be thought of as an edge-based measure to foil
attempts at intercepting communications in the core.87
Yet while encryption can empower users, it can also limit the use of Internet-
based content by curtailing users control over the applications running on their own
terminals. For example, Adobes Acrobat software facilitates the creation of e-
books,88 electronic versions of books that can be viewed on a computer screen.
While this technology can make books more accessible, copyright activists have
charged that it can also attenuate citizens rights of access to information by reducing
or eliminating fair dealing89 rights.
Such rights have historically been relatively easy to exercise (e.g., photocopying
by students) without prior restraint or self-enforcing limitations. Lawyers for
copyright holders might draft licensing terms that attempt to prohibit any non-
revenue-generating use of a work, but fair dealing rights override such terms. Lessig
contrasts the over-reaching of lawyers with self-enforcing rules like e-book
settings:
Now the over-reaching of an e-book that says, You can read this on a
Windows machine, but not on a Macintosh, is something more than bluster. It
is a set of controls with the power of mathematics behind itwe call that
encryptionand now these controls have the power of law to defend them
we call that the Digital Millennium Copyright Act.90
The American Digital Millennium Copyright Act91 contains anti-circumvention
provisions that make tinkering with such applications a criminal offence.92 Industry
86 Bruce W. Stratton, Data Security and Privilege on the Internet (1996) 12 C.I.P.R. 303.
87 A topic to which we return in Part II.C.3 in the context of ISPs and lawful interception.
88 See Adobe Systems Incorporated, eBooks Central, online: Adobe
285ff.
90 Lawrence Lessig, The Architecture of Innovation (2002) 51 Duke L.J. 1783 at 1797.
91 Pub. L. No. 105-304, 112 Stat. 2860 (1998) [DMCA].
92 See Pamela Samuelson & Suzanne Scotchmer, The Law and Economics of Reverse
Engineering (2002) 111 Yale L.J. 1575 at 1647. An interesting introduction to this important area,
which engages core values of copyright and freedom of expression, is an article by Jonathan Zittrain,
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
599
Canadas 2001 public consultation on digital copyright issues included the subject of
digital rights management systems,93 leaving Canadian anti-copyright activists
concerned that Canada might adopt its own DMCA-like regime.94
On the topic of server-side applications, another debate is raging over the extent
to which e-commerce business processes (which are generally embodied in server
software) can be patented.95 Perhaps the most controversial server-side issue,
however, is that of peer-to-peer media (primarily music) file-sharing. While rights
management mechanisms may be attenuating user freedom at the application layer,
peer-to-peer software has the potential to expand itthough this potential has not
gone unnoticed by what the peer-to-peer community calls the copyright industry.
The enormously popular Napster system was successfully destroyed by the Recording
Industry Association of America,96 and even its decentralized progeny (such as
Morpheus, KaZaA, and Gnutella),97 which do not feature centralized servers98 and
whose operators are obvious targets for litigation, are now being pursued.99
Nonetheless, client-side, peer-to-peer applications may represent the ultimate in user
empowerment. These applications can give file-sharers access to all the free music
they can download and turn each terminal into a mini-server.
Ironically, the compulsory tariff for the communication of musical works over the
Internet, which Canadian copyright collectivethe Society of Composers, Authors,
What the Publisher Can Teach the Patient: Intellectual Property and Privacy in an Era of Trusted
Privication (2000) 52 Stan. L. Rev. 1201.
93 Industry Canada & Canadian Heritage, Digital Copyright, supra note 61. In response to the
consultation, the Minister of Industry tabled Supporting Culture and Innovation: Report on the
Provisions and Operation of the Copyright Act (3 October 2002), online: Industry Canada
94 For instance, see the e-mail archives of the Canada DMCA Opponents Forum, online: Digital
Copyright Canada
95 Stephen J. Ferance, Debunking Canadas Business Method Exclusion from Patentability (2001)
17 C.I.P.R. 493; Richard Naiberg, Patent Protection for E-Commerce Inventions (2000/2001)
Internet and E-Commerce Law in Canada 17.
96 Napster and the much larger issue of copyright in music on the Internet are the topics of Alex
Colangelo, Copyright Infringement in the Internet Era: The Challenge of MP3s (2002) 39 Alta. L.
Rev. 891; W. Victor Tuomi, Music, Copyrights, and the Internet: The Copyright Board Chimes In
(2001) 18 C.I.P.R. 69. See also Michael D. Mehta, Don Best & Nancy Poon, Peer-to-Peer Sharing on
the Internet: An Analysis of How Gnutella Networks are Used to Distribute Pornographic Material
(2002) 1:1 C.J.L.T., online: Dalhousie University Electronic Text Centre
97 Online: StreamCast Networks
98 For a detailed discussion of the many incarnations of peer-to-peer file-sharing applications and
their legal status, see Tim Wu, When Code Isnt Law (2003) 89 Va. L. Rev. 679. Wu argues that
each successive design was a conscious response to copyright enforcement strategies.
99 See John Borland, U.S. Liability Looms over Kazaa CNet News.com (25 November 2002),
online: CNet News.com
Traders, Studios Spar in Court CNet News.com (2 December 2002), online: CNet News.com
600
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
and Music Publishers of Canada (SOCAN)has been pursuing for several years
before the Copyright Board, might be ineffective against peer-to-peer networks.100
The proposed tariffs scope has been shaped by two important decisions, one of
which was made by the Copyright Board and the other by the Federal Court of
Appeal.101 As it now stands, the tariff would apply to persons who post music files on
a server thereby making them available over the Internet.102 Even if individual
members of peer-to-peer file-sharing communities were considered to have posted
simply by allowing the file-sharing software to broadcast what is temporarily
available on the users hard drive for downloading by others, the transaction costs of
negotiating licences with each individual would be prohibitive. The Supreme Court
of Canada will soon decide on the legal status of the common ISP practice of
catching, whereby local copies of Internet content are stored on an ISPs server to
speed delivery to customers who later request that particular content. The Federal
Court of Appeal previously ruled that catching is not strictly necessary to ISPs
intermediary role, in which they benefit from an exception in the Copyright Act for
entities that merely provide the means of telecommunication necessary for another
person to communicate the work.103
1. Example: Overlay Software
The volume of valuable content available for free on the Web continues to
astound. What constitutes valuable content, of course, varies by user, but there is no
denying the benefits of free access to newspapers, for example. While some news
sites have always required paid subscriptions, and others require the user to endure a
privacy-sacrificing free registration process, most continue to be available at no
cost, as they have been since the dot-com boom days. These sites, however, tend to
be advertising-supported and as pressure increases to make Web media properties
profitable, advertising techniques have become ever more aggressive and annoying.
In response, entrepreneurs offer programs with names like Ad Blocker104 and
Junkbusters105 that allow savvy users to block banner and pop-up advertisements
from appearing on their screen during browsing sessions.
In the grey area in the middle of this advertising arms race is a type of client-side
application software known as overlay software. Overlay software, such as
100 A view suggested by Tuomi, supra note 96 at 90-91.
101 Re SOCAN Statement of Royalties, Public Performance of Musical Works 1996, 1997, 1998
(Tariff 22, Internet) (1999), 1 C.P.R. (4th) 417 (Copyright Board), reversed in part, [2002] 4 F.C. 3,
215 D.L.R. (4th) 118 (C.A.), leave to appeal and cross-appeal to S.C.C. granted, [2002] S.C.C.A. No.
289 (QL) [Tariff 22].
102 Michael Koch, Only in Canada, You Say? Perhaps Not: Federal Court of Appeal Rules on
Internet Copyright (2002/2003) 3 Internet and E-Commerce Law in Canada 25 at 27.
103 R.S.C. 1985, c. C-42, s. 2.4(1)(b).
104 Online: Ad Blocker
105 Online: Junkbusters
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
601
Gator,106 is marketed as online companion software and claims to enhance the
browsing process by automating functions such as filling out forms and entering
passwords. Marketed is perhaps not the right word: the software appears more often
to be acquired unintentionally by downloading other free software programs, such as
the peer-to-peer file-sharing program KaZaA. Bundled with the Gator eWallet,
however, comes another Gator program, called OfferCompanion, touted as your
direct link to some of the Webs most valuable offers. OfferCompanion is, as the
Gator Web site acknowledges, ad-supported software. While the user browses the
World Wide Web, the software communicates with Gators servers and through the
magic of contextual advertising, pop-up advertisements appear on the users screen
that have been paid for by Gators advertiser customers. The problem, in the view of
the publishers of several of the Webs premiere commercial media Web sites, is that
these advertisements often appear directly over top of advertisements that would
otherwise appear on the visited Web site.
The Washington Post Company, the New York Times Company, Dow Jones, and
other media companies brought an action against Gator107 alleging trademark and
copyright infringement, among other things.108 The plaintiffs were granted an
injunction restraining Gator from placing advertisements over their Web pages,109 and
an undisclosed settlement followed.110 The Web publishers argued that Gator
interfered with their Web pages, in which they have made significant investments,
without their permission. Gator countered that the software does no such thing, that
the plaintiffs Web sites were not altered in any way. Rather, the sites just appeared
differently on the screens of those users running the Gator software. Ottawa
intellectual property lawyer Eric Smith has taken the publishers position, arguing that
they may have had good claims in copyright infringement, trademark infringement,
and unfair competition under Canadian law.111
Regardless of whether the publishers would have succeeded in making out such
claims, there is a larger issue at stake. This kind of dispute has implications for the
extent to which Internet users can control the way Web content appears on their
computer screens. The Gator software does not actually alter the content of Web
pages. It does not achieve the same result as a hacking, whereby the content of the
106 Online: Gator
107 Washingtonpost.newsweek Interactive, LLC. v. Gator Corporation, [2002] U.S. Dist. LEXIS
20881 (E.D. Va.).
108 Stefanie Olsen, Publishers Sue Gator over Pop-Ups CNet News.com (27 June 2002), online:
CNet News.com
109 Stefanie Olsen, Judge: See Ya Later, Gator CNet News.com (12 July 2002), online: CNet
News.com
110 Stefanie Olsen, Court says Gator-style Ads Are Legal CNet News.com (1 July 2003), online:
CNet News.com
111 Eric J. Smith, Contextual Advertising: The Case Against Theftware (2002/2003) 3 Internet
and E-Commerce Law in Canada 33.
602
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
site as it sits on the providers server is changed.112 Only the way a particular page is
displayed on a particular users screen is altered. Other overlay software products
offer functionality such as highlighting key words, links to reference sources (often
paid advertisements), and annotations written by other surfers. In fact, Internet users
can dramatically alter the appearance of every Web page they visit by simply
changing their browser settings so as not to display graphics (thus excluding most
advertisements). Accessibility features in Microsoft operating systems allow
similarly extensive alterations, for instance, by displaying text in very large size for
the benefit of the visually impaired. Users can further control their Web experience by
blocking cookies and Java scripts (mini-programs which can be delivered and
executed along with Web content) if they know how. The difference in the Gator
case, of course, is that the software permits such modification on a mass scale, and
the distributor profits from the modifications.
The legal right of consumers to avoid advertisements on free media is not a new
issue; in fact, it is also playing out in the case of digital television recorders (such as
ReplayTV and TiVo).113 These devices allow viewers effectively to skip over the
advertisements on television programs. Other devices can mute the sound or change
the channel when advertisements are played.114 The president of Turner Broadcasting
caused a furor in mid-2002 when he suggested (in the words of one news article) that
people who watch television without commercials were stealing from entertainment
producerswith possible exceptions made for folks who need to use the
bathroom.115 The same dividing line between consumer self-help measures to avoid
advertising and for-profit attempts to divert eyeballs (or consumer attention) is
being explored in both cases.
As the overlay software example suggests, the scope of rights relating to the
application layer remains unclear, yet these rights are critical to the ability of users at
the edge to control what the Internet does for them. While copyright issues relating to
the protection of valuable material on-line dominate at the content sublayer,
copyrights role at the application layer is to protect the applications used to access
that content. In both realms, fair dealing rights claimed or historically enjoyed by
112 Canada proved to be the home of the most notorious Web hacker to date, mafiaboy, a
Montreal teen who intermittently crippled the Web sites of Amazon, CNN, Dell, eBay, and Yahoo!
from 7-15 February 2000 by means of a distributed denial of service attack in which Web servers were
flooded with so many requests for data that they were effectively clogged. He was charged under
subsections 342.1(1) (unauthorized use of computer) and 430(1.1) (mischief in relation to data) of the
Criminal Code (R.S.C. 1985, c. C-46) and sentenced on 12 September 2001 to eight months detention
plus one year probation (R. c. M.C., [2001] J.Q. no. 4318 (C.Q. jeun.) (QL)).
113 Online: Replay TV
114 See e.g. the Ad Zapper, online:
115 Michael Freedman, Zapper War: TV Producers in Quest to Outlaw Device that Allows
Consumers to Skip Ads ABCNews.com (27 June 2002), online: ABCNews.com
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
603
consumers are being challenged by the strong exclusionary claims of corporate
copyright holders, which are in turn backed up by self-enforcing technical constraints.
Applications like overlay software and peer-to-peer file-sharing programs allow
the user to override the control of content that the entertainment industry enjoys in
traditional media by virtue of the existence of gatekeepers like advertising-
supported broadcasters. In the absence of such central points of control, however,
strategies such as the limitation of client-side applications (e.g., e-book viewers and
browsers) are being pursued by those seeking to regain hegemony, often through
compliant intermediaries like computer manufacturers. Unfortunately, few users are
aware of these strategies.
The overlay software case also raises questions about the individual users ability
to control which applications run on his or her PC. Surreptitiously installed programs
like Gator that produce commercial benefits for their sponsors (in this case, browsing
data prized by advertisers) present privacy concerns. On the other hand, malicious
applications like viruses and Trojan horses present all-too-common risks of property
damage and data loss. What is the nature of the users legal rights in this regard?
Does the user implicitly cede some control over his or her machine by plugging it into
the Internet? If so, are there general rules that mediate all Internet participants
vulnerabilities?
These questions are only now beginning to be posed, thanks to increasing
awareness of the importance of the application layer to the overall Internet
experience. The inherently international nature of the application layer (as opposed
to, for example, the physical layer) will complicate the resolution of these issues. That
said, if one jurisdiction can force the producers of a mass-market software application
to give effect to a particular policy choice, such as limiting the strength of data
encryption features, then that decision can have a global effect. This problem further
illustrates the importance of user control over application software and of keeping
applications conceptually separate in legal and policy analysis.
C. Operational Layer
There has been relatively less legislative and adjudicative activity to date directly
relating to the operational layer, but the issues are no less important. Legal issues at
the operational layer are often referred to as matters of Internet governance,
although the scope of this term is itself a matter of some debate. To those who see
broad policy implications in the control of the essential centralized resources and
functions, the term is almost akin to Internet regulation. For others who seek to
keep the operational realm free of politics (as if that were possible), these are matters
of mere technical coordination. The sources of legal and quasi-legal authority in
this realm are unusual, and in most cases, institutions and rules are still very much in
the formative stage.
In addition to the familiar domain name issue, several lesser-known operational
layer topics are introduced below. These issues are examined in the context of three
604
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
sublayers: (1) centralized resources and functions; (2) standards and protocols; and
(3) ISP functions. The example of e-mail service brings these three groups of
elements together and illustrates the impact that they have on the way the Internet
works as a whole.116
1. Centralized Resources and Functions Sublayer
The elements of the centralized resources and functions sublayer include network
identifiers (i.e., names, numbers, and addresses), the DNS, and the RSS. During the
early, non-public internetworking era when access was restricted to U.S. government
agencies and contractors, and later scientists at universities, these resources and
systems were effectively self-governedwhich is to say that authority over them was
informal and often wielded by trusted individuals known in the network community.
As the Internet expanded, and particularly when commercial ISPs joined in, the
potential scarcity of these resources slowly started to prompt questions about their
governance.
The first and most prominent of such clashes was the still unresolved conflict
between trademarks and domain names. The friction between trademark systems and
the DNS has been extensively treated by legal scholarship around the world,117 and in
Canada.118 Trademark systems can accommodate multiple simultaneous uses of
similar brands and names worldwide, while the DNS does not, or at least has not been
permitted to do so by the powerful global trademark lobby.119 While domain names
can be thought of as elements of the Internets centralized resources and functions,
domain names can implicate a much broader set of legal issues. As Teresa Scassa has
noted, questions surrounding the right to use particular domain names (among other
expressive on-line activities, like Web site design) raise serious concerns about the
116 See Part II.C.4, below.
117 For some of the more insightful domain name/trademark articles, see Dan L. Burk, Trademarks
Along the Infobahn: A First Look at the Emerging Law of Cybermarks (1995) 1 Richmond J.L. &
Tech. 1; Robert Shaw, Internet Domain Names: Whose Domain is This? in Kahin & Keller, supra
note 37, 107.
118 Teresa Scassa, Intellectual Property on the Cyber-Picket Line: A Comment on British Columbia
Automobile Assn. v. Office and Professional Employees International Union, Local 378, Case
Comment (2002) 39 Alta. L. Rev. 934 [Scassa, Cyber-Picket Line]; Bradley J. Freedman & Robert
J.C. Deane, Trade-marks and the Internet: A Canadian Perspective (2001) 34 U.B.C. L. Rev. 345;
Chad Mitchell Bayne, Domain Names: A Canadian Perspective (2000) 17 C.I.P.R. 31; Lisa Katz
Jones, Trademark.com: Trademark Law in Cyberspace (1999) 37 Alta. L. Rev. 991; Jonathan E.
Moskin, Canada and the Future of Internet Governance (1999) 15 C.I.P.R. 247; Robert M. Frank,
Cybernames: Domain Name Issues and Conflicts in Cyberspace (1996) 12 C.I.P.R. 245; Andrea F.
Rush, Internet Domain Name Protection: A Canadian Perspective (1997) 11 I.P.J. 1.
119 An argument made in Milton L. Mueller, Ruling the Root: Internet Governance and the Taming
of Cyberspace (Cambridge: Massachusetts Institute of Technology Press, 2002) at 231 [Mueller,
Ruling the Root].
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
605
balance being struck between monopolistic intellectual property rights and the
fundamental right of freedom of expression.120
The most significant new legal regimes created in the operational context to date
are dispute resolution systems for the allocation of domain names. Most are modelled
after the Uniform Domain Name Dispute Resolution Policy (UDRP) of ICANN.121
These processes, depending on who you ask, either help the trademark industry
prevent individuals and protest groups from using whatever characters they like in
their domain names or help to rid the Internet of cybersquatting: the opportunistic
registration and auctioning of attractive names, particularly those implying an
association with well-known corporations.
Domain name law is now a field unto itself,122 but the process of formalizing
Internet governance is deserving of further mention. In the mid-1990s, outsiders
began examining the power over domain names that was held by individuals and
certain institutions. The U.S. National Science Foundation initially had executive
authority. Later,
the U.S. National
Telecommunications and Information Administration, which administered operating
contracts with some individuals with considerable influence in the area. The reform
process that culminated in the creation of ICANN constituted an unprecedented
process of codifying unwritten Internet law.123
responsibility was
assumed
by
Before ICANN, people and agencies often held decision-making power over key
elements of the Internets operational infrastructure only by historical accident. In
other cases, government contracts that began as obscure procurements for a U.S.
academic research network became the root of enormous power (and in some cases
wealth) in the commercial era.124 ICANN is intended to serve as the international,
non-profit, non-governmental, industry self-regulating body for the Internets key
centralized resources and functions, including the DNS. While ICANN continually
insists that it is only engaged in technical policy-making for a narrow set of
coordinated Internet functions, others view it as the full-fledged regulator of the DNS
120 Scassa, Cyber-Picket Line, supra note 118 at 934.
121 Online: ICANN
Deane, The Uniform Domain Name Dispute Resolution Policy: A Practical Guide (2002) 1:1
C.J.L.T., online: Dalhousie University Electronic Text Centre
Cybersquatters Not Welcome: A Review of Domain Name Dispute Resolution Procedures (2001)
18 C.I.P.R. 103.
122 A good source for current information is the online companion to Ellen Rony & Peter Rony, The
Domain Name Handbook: High Stakes and Strategies in Cyberspace (Lawrence, Kan.: R&D Books,
1998), online:
123 See A. Michael Froomkin, Wrong Turn in Cyberspace: Using ICANN to Route Around the
APA and the Constitution (2000) 50 Duke L.J. 17.
124 The best place to start an investigation into these fascinating arrangements is the U.S.
government policy statement known as the White Paper. See U.S. Department of Commerce,
Management of Internet Names and Addresses (5 June 1998), online: U.S. Department of Commerce
606
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
and possibly other operational layer functions.125 ICANNs major accomplishments to
date include the promulgation of the UDRP and the admission of seven new, generic
Top-Level Domains (gTLDs) to the existing .com, .net, and .org: .aero,
.biz, .coop, .info, .museum, .name, and .pro. This was accomplished,
however, only after an acrimonious two-and-a-half-year process.126
At the international level, the most sensitive Internet governance issue is the U.S.
governments continuing role as the source of legal authority for the management of
the top levels of the centralized resources and functions.127 The European Union has
been the most outspoken critic of the U.S. governments continuing involvement,128
and even the latters own 1998 policy paper recommended that it withdraw
completely.129 The proper role of governments in this sphere, and particularly that of
the U.S. government, are major unresolved issues in international Internet policy.130
2. Standards and Protocols Sublayer
This layer consists mainly of the TCP/IP suite, the language spoken by all
computers on the Internet. It has been suggested that while edge code is inherently
diverse, core code, like the TCP/IP suite, is inherently uniform. This difference
emphasizes the importance of understanding the different patterns of control over the
two types of code. In many obscure, technical ways, the architecture of core code
(and in particular the TCP/IP protocol suite) limits what is possible on the Internet.
Counterbalancing the tremendous power of this kind of code is the historical fact that
the TCP/IP suite is in the public domain; that is, the TCP/IP suite is typically not the
property of any one vendor. Since adopting these standards and protocols is
essentially compulsory in order to participate in the Internet, policy-makers must be
aware of the institutions and politics of this obscure realm. Unfortunately, scholars are
only in the early stages of exploring these matters, so the legal literature is quite thin.
125 Milton Mueller, ICANN and Internet Governance: Sorting Through the Debris of Self-
Regulation (1999) 1 Info 497; Jonathan Weinberg, ICANN as Regulator (Paper presented at the
29th Research Conference on Information, Communication, and Internet Policy, October 2001)
[unpublished], online: Cornell University
126 See Mueller, Ruling the Root, supra note 119 at 201ff.
127 See Kim G. von Arx & Gregory R. Hagen, Sovereign Domains: A Declaration of Independence
of ccTLDs from Foreign Control (2002) 9 Rich. J.L. & Tech. 4.
128 See European Parliament Resolution on the Commission Communication to the Council and
the European Parliament on The Organisation and Management of the InternetInternational and
European Policy Issues 1998-2000 (COM(2000) 202C5-0263/20002000/2140(COS)) in EC,
Sitting of Thursday, 15 March 2001, [2001] O.J. C. 44-343/198 at 286.
129 U.S. Department of Commerce, supra note 124.
130 This is the subject of Craig McTaggart, The ENUM Protocol, Telecommunications Numbering,
and Internet Governance Cardozo J. Intl. & Comp. L. [forthcoming in 2004].
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
607
The primary Internet standards development body is the Internet Engineering
Task Force (IETF).131 The IETF was founded in 1996 by a very small group of like-
minded engineers who shared a common interest in developing interoperable TCP/IP
networks. By the mid-1990s, meeting attendance began to exceed two thousand
people, the vast majority of whom were representatives of Internet industry
equipment vendors and software developers. The term standards-setting is not,
however, an entirely accurate way to describe what the IETF does. The IETF is a
forum in which operational and architectural ideas are proposed and experimented
with. Ideas that meet with some measure of approval from interested members are put
forward as standards that equipment vendors and networks are free to adopt or ignore:
The IETF does not impose standards.132 Philip Weiser describes the open standards
bargain as follows:
The Internets openness created a virtuous cycle where members of the
Internet community continued to improve upon its basic architecture by
adding new functionalities that were placed in the public domain, thereby
making the Internet a more valuable network. On the supply side, a culture
emerged whereby developers would work with one another and rely on open
standards rather than compete with one another to establish the basic
architecture that supports the Internet. Because trusted standard-setting
organizations adopted these key standards and made them open, developers
did not have to worry about these standards being ignored and defeated,
thereby undermining the value of any applications built off of those
standards.133
It is essential that the Internet industry continue to see the value in open standards and
interoperability. The alternative is to have multiple non-interoperable, non-
interconnected systems, somewhat akin to the state of the computer industry before
TCP/IP allowed communication between equipment of different types running
different software. In the clever words of Sharon Eisner Gillett and Mitchell Kapor
(co-founder of Lotus and the Electronic Frontier Foundation), Interoperability is like
Tinkerbell: it only works if everyone believes in it.134 The same could be said for the
Internet generally, where interoperability is based on open standards like those of the
IETF.
131 Online: IETF
132 See S. Bradner, RFC 2026: The Internet Standards Process (Rev. 3) (October 1996), online:
IETF
official technical documentation series. There are several different types of RFCs, of which Internet
standard is only one. For more on the IETF, see IETF Secretariat & G. Malkin, RFC 1718: The Tao
of the IETF: A Guide for New Attendees of the Internet Engineering Task Force (November 1994),
online: IETF
133 Philip J. Weiser, Internet Governance, Standard Setting, and Self-Regulation (2001) 28 N. Ky.
134 Sharon Eisner Gillett & Mitchell Kapor, The Self-Governing Internet: Coordination By Design
L. Rev. 822 at 826-27.
in Kahin & Keller, supra note 37, 3 at 16.
608
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
that has been stable (an engineering
While the members of the IETF can theoretically change the way the Internet
works by amending its core code, change tends to happen rather slowly. New
standards and protocols must be implemented by ISPs one by one. They cannot be
imposed Internet-wide all at once. This point is illustrated by the time that it is taking
to upgrade the TCP/IP suite from its current version, IPv4, to a new version,
IPv6,
term meaning ready for
implementation) since 1998.135 This upgrade is considered crucial, both for avoiding
a shortage of IP addresses and enabling new security and routing features.136
Implementation has been very slow, however, since IPv4 appears to be serving the
needs of the internetworking community well enough for now (at least in North
America, where IP addresses are plentiful). The current slump in the Internet industry
is making investment in new equipment that is IPv6-enabled even slower. The
transition to IPv6 will not likely happen until enough large (i.e., American) ISPs hear
their customers demanding it.
The Internets remarkable decentralization, which empowers its edges, can also
be a weakness by preventing its constituent networks from taking Internet-wide
collective action, even when the technical experts insist that such action is necessary.
This decentralization may become more of a problem as expectations of the Internet
continue to grow (e.g., that it be able to carry television-quality video). These
problems will, in turn, likely have the effect of increasing the profile of the IETF and
its processes.
3.
ISP Functions Sublayer
Standards and protocols are only useful if they are adopted by ISPs, which
perform the actual transmission of Internet traffic. ISPs come in all sizes, from local
retail dial-up access providers, to large institutions like universities, to wholesale
backbone carriers that aggregate traffic and that, at least notionally, connect
everything to everything else. The sum of the myriad interconnections among ISPs
can perhaps be said to be the Internet.
ISPs and Internet applications have enjoyed something of a special status under
Canadian law to date.137 Of all the parties involved in making the Internet happen,
ISPs would seem to be the easiest target for regulation by domestic authorities
because they play a gatekeeper role and have operations, assets, and revenues within
at least one courts jurisdiction. However, ISPs frequently argue that they have no
online: IETF
136 To learn more, see the Web site of IPv6.org, a self-organized Internet technical community group
135 See S. Deering & R. Hinden, RFC 2460: Internet Protocol, Version 6 (IPv6) (December 1998),
which promotes the transition to IPv6, online: IPv6
137 See generally Andrew Bernstein & Rima Ramchandani, Dont Shoot the Messenger!: A
Discussion of ISP Liability (2002) 1:2 C.J.L.T., online: Dalhousie University Electronic Text Centre
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
609
knowledge of what customers use their Internet connections for, and thus should not
be held liable for harm caused by those uses.
The ISP industry, led by the Canadian Association of Internet Providers
(CAIP),138 has been very successful in convincing decision-makers to treat ISPs as
mere conduits for others activities. Consider, for example, the exemption from
copyright infringement liability recognized by the Copyright Board (and affirmed by
the Federal Court of Appeal) in the Tariff 22 proceedings.139 The regulatory status of
the ISP under the Telecommunications Act140 and Broadcasting Act141 was clarified
somewhat by the CRTC in its feel-good New Media decision (taken at the height of
dot-com mania), in which the commission acceded to public pressure to take a
hands-off approach.142
The most potentially controversial aspect of the relationship between ISPs and
their customers regards personal privacy. Ethan Katsh defines privacy as being able
to control information about oneself.143 In the Internet environment, this includes
ones on-line identity and habits. ISPs have the technical ability to monitor and record
every move that their customers make on the Internet. Ian Kerr has suggested that
ISPs may stand in a fiduciary relationship to their customers because ISPs are in a
position to learn highly personal information about their clients and sometimes must
decide in what circumstances and to whom this information should be revealed.144
Referring to the specific case of an ISP aiding a securities-related investigation,
Michael Geist has noted, The ISPs ability to cooperate under circumstances
mandated by legal necessity suggests that they may possess greater access to data
than they might have the public believe.145
Indeed, this access is precisely why the federal department of justice is currently
proposing measures to facilitate law enforcement interception and acquisition of e-
mail messages stored on ISP servers.146 An international movement in this direction
138 Online: CAIP
139 Supra note 101.
140 S.C. 1993, c. 38.
141 R.S.C. 1985, c. B-9.
142 New Media (17 May 1999), Telecom Public Notice CRTC 99-14 and Broadcasting Public
Notice CRTC 1999-84, online: CRTC
143 M. Ethan Katsh, Law in a Digital World (New York: Oxford University Press, 1995) at 228.
144 Ian R. Kerr, The Legal Relationship between Online Service Providers and Users (2001) 35
Can. Bus. L.J. 419.
145 Geist, Internet Law in Canada, supra note 51 at 89.
146 See Department of Justice Canada, Industry Canada & Solicitor General Canada, Lawful
AccessConsultation Document (25 August 2002) at Interception of E-Mail, online: Department
of Justice Canada
610
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
was underway prior to the events of 11 September 2001,147 but this initiative is now a
part of the broader federal strategy embodied in the Anti-terrorism Act.148 One
Canadian court has explicitly recognized that Internet e-mail ought to carry a
reasonable expectation of privacy, raising the bar for justifying lawful interception.149
Civil liberties and privacy advocates, however, remain concerned that these initiatives
are increasing law enforcement power beyond what can be justified in a free and
democratic society.150
While most Canadian ISPs are members of CAIP and endorse the CAIP Privacy
Code,151 this code is voluntary and each member is free to define how it subscribes to
each principle. On the other hand, ISPs are subject to the Personal Information
Protection and Electronic Documents Act152 and the Privacy Commissioner of Canada
has already investigated several complaints relating to ISP business practices.153
Internet privacy,154 of course, is just a small part of the much larger (and fast-growing)
field of privacy law.155 As noted at the beginning of this part, privacy is one of the
issues pervading all of the layers in this typology. Space does not permit a complete
examination of privacy issues in this paper. Nevertheless, a case from the Privacy
Commissioner serves as the introduction to an in-depth examination of issues at the
operational layer surrounding e-mail service.
147 See especially Council of Europe, Convention on Cybercrime, 23 November 2001, Eur. T.S. 185,
41 I.L.M. 282, online: Council of Europe
148 S.C. 2001, c. 41.
149 R. v. Weir (1998), 213 A.R. 285, 59 Alta. L.R. (3d) 319 (Q.B.), affd (2001), 281 A.R. 333, 156
C.C.C. (3d) 188 (C.A.).
150 See e.g. Lisa Austin, Is Privacy a Casualty of the War on Terrorism? in Ronald J. Daniels,
Patrick Macklem & Kent Roach, eds., The Security of Freedom: Essays on Canadas Anti-Terrorism
Bill (Toronto: University of Toronto Press, 2001) 251.
151 Canadian Association of Internet Providers, CAIP Privacy Code, online: CAIP
152 S.C. 2000, c. 5 [PIPEDA].
153 See e.g. Privacy Commissioner of Canada, Unsolicited E-Mail from an Internet Service Provider
(3 July 2001), online: Privacy Commissioner of Canada
154 See also James A. Fontana, The Law of Search and Seizure in Canada, 5th ed. (Markham, Ont.:
Butterworths, 2002), c. 16, Technical and Electronic Surveillance, c. 23, Computer-Related
Searches; Robert W. Hubbard, Peter DeFreitas & Susan Magotiaux, The Internet: Expectations of
Privacy in a New Context (2002) 45 Crim. L.Q. 170; John G. Boufford, Privacy on the Information
Highway (1998) 47 U.N.B.L.J. 219.
155 See e.g. Colin H.H. McNairn & Alexander K. Scott, Privacy Law in Canada (Markham, Ont.:
Butterworths, 2001); Stephanie Perrin et al., The Personal Information Protection and Electronic
Documents Act: An Annotated Guide (Toronto: Irwin Law, 2001); Teresa Scassa, Text and Context:
Making Sense of Canadas New Personal Information Protection Legislation (2000/2001) 32 Ottawa
L. Rev. 1; Barbara McIsaac, Rick Shields & Kris Klein, The Law of Privacy in Canada, looseleaf
(Toronto: Carswell, 2000); Michael Power, Bill C-6: Federal Legislation in the Age of the Internet
(1999) 26 Man. L.J. 235.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
611
4. Example: E-Mail Service
One issue that seems bound to lead to litigation and perhaps even legislation in
the future is that of e-mail service. Perhaps without even noticing, Canadians have
come to depend on Internet e-mail as a major, or in some cases primary, means of
communication. Even important documents like job applications are routinely sent by
e-mail (and e-mail alone) today. Many universities allow graduating students to keep
the addresses they had as students, recognizing the value of having a direct link to the
hearts and minds of future donors. Use and misuse of work e-mail accounts has
given rise to a host of problems of its own.156 Now that many people view their e-mail
address as an essential part of their identity, and their e-mail account as an essential
means of communication, we are starting to see how high expectations for e-mail
service can be.
E-mail is an application, and one of the Internets most popular at that, but it is
different in one key respect from other applications such as Web browsers. E-mail
requires the services of an intermediary to send, receive, and store e-mails for users
(i.e., to operate local mail servers). This intermediary is normally an ISP, and because
of the necessity of their intervention, e-mail is best thought of as an operational layer
issue.
A Privacy Commissioner case involving the withholding of e-mails illustrates both
the degree to which some people are becoming dependent on e-mail, and how unclear
its legal status remains. The commissioners summary of the complaint sets the stage:
An individual complained that her internet service provider (ISP), by
continuing to take in and store her e-mails while her account was under
suspension and by withholding them from her pending payment of arrears,
had improperly used her personal information without her knowledge and
consent for a purpose other than that for which it had been collected.157
After the second time the ISP suspended the subscribers Internet service account, she
contacted the ISP and made a reference to canceling her account, but does not appear
to have unequivocally communicated that intention. The ISP continued to collect e-
mail messages on her behalf, but would not give her access to messages until her
arrears were paid. According to the commissioners findings, she never paid those
arrears, but subsequently became aware that messages to that account were not
156 Kevin Coon & Jonathan Cocker, Legal Issues of E-Mail and Internet Access in the Workplace
(2000/2001) 1 Internet and E-Commerce Law in Canada 81; Charles Morgan, Employer Monitoring
of Employee Electronic Mail and Internet Use (1999) 44 McGill L.J. 849; Holly L. Rasky, Can an
Employer Search the Contents of Its Employees E-Mail? (1998) 20 Advocates Q. 221.
157 Internet Service Provider Accused of Withholding E-Mails Sent to Suspended Account (28
August 2002) Privacy Commissioner of Canada Finding 66 at para. 1, online: Privacy Commissioner
of Canada
(complaint well-founded)
[Withholding E-Mails].
612
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
bouncing back to senders as undeliverable, and with persistence she was able to
retrieve those messages.158
With respect to the subscribers complaint that the ISPs actions contravened
PIPEDA, the commissioner found that the governing service agreement did not give
her any notice of how the service provider would handle her incoming e-mails during
a suspension period and thus could not have formed a sufficient basis for her
knowledge and consent regarding such practices. Finally, the commissioner
concluded, Nor could the ISPs continued absorption, storage, and withholding of e-
mails pending payment of arrears be in any reasonable sense deemed uses consistent
with the purpose for which the company ordinarily collected e-mails on behalf of its
clients.159 The ISP was found to have used the subscribers personal information
without her consent for purposes other than that for which it had been collected; thus,
the ISPs actions contravened PIPEDA. However, the Privacy Commissioner
recognized the validity of the contract between the subscriber and the ISP and noted
that the ISP had since amended its standard agreement to give better notice of this
practice, thereby complying with the legislation.
The commissioner, however, did not stop there and neither did the subscriber.
The summary continues, under the heading Further Considerations:
The Commissioner remained concerned about the implications of storing and
withholding potentially important messages without informing the intended
recipient of their existence or the sender of their non-delivery. The practice
falsely leads the sender to believe that the message has gone through
unimpeded. The Commissioner recommended, in the interests of best practice,
that the ISP immediately cease collecting, storing, and denying access to, e-
mails addressed to holders of accounts under suspension and adopt instead the
practice of deflecting such e-mails back to the senders with notification to the
effect that the messages could not be delivered.160
This non-binding statement illustrates the commissioners personal concern for the
sanctity of e-mail, but it is difficult to see how this scenario would remain a privacy
issue (and thus within his jurisdiction) if a service agreement that gives adequate
notice of the practice does not contravene PIPEDA.
Perhaps emboldened by the commissioners attempt to urge the development of a
normative duty to bounce, the subscriber soon after filed a well-publicized civil
action in the Federal Court seeking compensatory damages in the amount of $80,000
and punitive damages of $30,000.161 One of the hostage emails was allegedly a job
158 Ibid. at para. 4.
159 Ibid. at para. 11.
160 Ibid. at para. 15.
161 Carter v. Inter.net Canada Limited (11 October 2002), Toronto T-1745-02 (F.C.T.D.). See also
Evan Hansen, Who owns your E-mail? CNet News (29 October 2002), online: CNet News
message for Internet service providers: Keep your hands off my e-mail).
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
613
offer, which the subscriber lost because the sender assumed she was not interested in
the position since she did not reply, as was her custom, to that particular message.
While section 16(c), the remedies section of PIPEDA, stipulates that the court may
award damages to the complainant, including damages for any humiliation that the
complainant has suffered, whether this kind of economic loss will be found to have
been a foreseeable consequence of the ISPs impugned actions (the use of the
subscribers personal information without her consent for purposes other than that for
which it had been collected) remains to be seen.162 If the subscriber can neither prove
that she instructed the ISP in unequivocal terms to terminate her account nor point to
conduct consistent with a belief that she had done so (e.g., sending out the customary
change-of-address message), her chances of success seem slim.
The commissioners decision demonstrates that so long as ISPs give adequate
notice of, and so long as subscribers consent to, ISP business practices that may be
considered privacy-invasive, PIPEDA does not prohibit such practices. Whether
subscribers actually read their service agreements is another matter. If they did, they
might be surprised at what they would find. What is the ISPs liability when an e-mail
sent by a correspondent is never received at all instead of being withheld by the ISP
(as almost everyone has experienced at least once)? The terms of service binding the
largest single group of residential Internet access customers in Canada, those of the
Sympatico network, are typical:
You expressly understand and agree that: (a) the Sympatico network and the
Services are provided on an as is and as available basis and (b)
Sympatico specifically makes no warranties that the Sympatico Network or
any of the Services, including any content, information, products or services
obtained from or through the use of the Sympatico Network or the Services,
will be provided on an uninterrupted, timely, secure or error-free basis or that
such services or the results derived therefrom will meet your requirements or
expectations.163
Imagine a disclaimer applying to any other service as important (in many
peoples eyes) as Internet access which states that the service will only be provided
as available and is not warranted to meet your requirements or expectations.
Surely customers are entitled at least to expect the delivery of e-mail messages.
However, both courts164 and the CRTC165 have found the Internet access market to be
162 The commissioner nevertheless found the impugned actions to be standard industry practice
(Withholding E-mails, supra note 157 at para. 14).
163 Sympatico Inc., Sympatico Network Terms and Conditions (last updated 1 April 2003), s. 19,
online: Sympatico
have converted this passage from upper-case to sentence-case letters.
164 In 1267623 Ontario Inc. v. Nexx Online Inc. (1999), 45 O.R. (3d) 40 at 50, 46 B.L.R. (2d) 317
(S.C.J.), which is currently the leading Canadian case dealing with spam and netiquette, Wilson J.
noted that a subscriber whose account has been terminated by an ISP could simply find a new one.
Space does not permit me to consider the problem of spam (or unsolicited commercial e-mail), other
than to note that as of July 2002, spam was thought to constitute over one-third of all e-mail traveling
614
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
competitive, and one presumes that the market will sort out which ISPs provide good
service.
In the telephone industry, by contrast, companies providing basic service have to
meet stringent quality standards.166 It may only be a matter of time before consumers
call for ISPs to be similarly regulated. Even before Rogers Communications bungled
the transition of its high-speed Internet customer base from @home.com e-mail
addresses to @rogers.com e-mail addresses in late 2001,167 a group of disgruntled
Toronto area customers attempted to bring a class action against Rogers for alleged
cable Internet service interruptions.168 For its part, the Residential Broadband Users
Association (RBUA), a consumers group, stayed out of the dispute, preferring to
maintain dialogue with Rogers in the hopes of improving service for all customers.169
While some service quality problems are doubtless the fault of ones ISP, the
reality is that Internet performance depends on many different factors, including the
service provided by the dozens of different ISPs that might be involved in routing any
given e-mail from one side of the world to the other and the availability of key
centralized resources and functions, such as the RSS. Even if Parliament or the CRTC
wanted to regulate,170 ISPs would protest that they simply cannot control the Internet
at large and therefore cannot make any promises in respect of its performance.
Viewed from this perspective, one might think it is a minor miracle that the Internet
works at all, and perhaps it is. The availability of the content layer, and the operation
of the application layer underneath it, both depend on the performance of the
operational layer.
This is why the least-known realm of the Internet, the largely invisible
operational layer, may present the most important legal and policy challenges for the
future. Operational layer elements have either avoided regulatory attention, as in the
case of the standards and protocols sublayer, or been the subject of experimental and
hands off policies to date, as in the cases of the centralized resources and functions
over the Internet (Robert Lemos, Youve Got Spam, and More Spam CNet News (29 August 2002),
online: CNet News
165 Regulation Under the Telecommunications Act of Certain Telecommunications Services Offered
by Broadcast Carriers (9 July 1998), Telecom Decision CRTC 98-9, s. 19, online: CRTC
166 CRTC Creates New Quality of Service Indicators for Telephone Companies (9 April 2001),
Telecom Decision CRTC 2001-217, online: CRTC
167 The transition was necessitated by the bankruptcy of U.S.-based Excite@Home, of which
Rogers Communications was a shareholder. See Patrick Brethour, Rogers E-Mail Clients Drag Feet
on Switch Away from At Home The Globe and Mail (30 November 2001) B1.
168 This was the subject of the Kanitz case, supra note 85, in which Rogers was successful on a
motion to have a proposed class action stayed on the grounds that the user agreement stipulated that
any disputes arising under it were to be referred to, and determined by, arbitration alone.
169 RBUA, News Release, Statement on class action lawsuit brought against Rogers Cable (23
July 2001), online: RBUA
170 On the question of the CRTCs jurisdiction over Internet services, see Part II.D.2, below.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
615
and ISP functions. Since these elements serve as the glue that holds the rest of the
Internet together, the current obscurity of their governance seems out of step with
their significance to the Internet at large. The operational layers components are
essential to all uses of the Internet, yet we know the least about who controls these
elements and what legal obligations apply or should apply to them.
Operational layer issues also highlight the jurisdictional puzzles that can beset
Internet-related analysis. In the case of e-mail, for example, the jurisdiction in which
the defendant is located would depend on whether the e-mail service provider is a
local ISP or a Webmail provider. Webmail providers can theoretically be located
anywhere in the world (the popular service hotmail.com, for example, is operated by
Microsoft Corporation and its central servers are located in Washington State). The
physical location of the server is also relevant to the question of whether SOCANs
proposed Internet music tariff would apply to a given communication of a work over
the Internet. Physical location also plays a role in determining whether telephone calls
over the Internet are subject to universal service subsidy charges, as explained in the
next section.
The novelty of the problems found at the operational layer, combined with the
layers international scope, can be expected to continue stretching traditional legal
regimes to the limit. The operational layer has already spawned sui generis regimes
such as ICANN. Among areas to watch in the future, the international legal status of
the IETF and other non-governmental stewards of standards and protocols will be
particularly interesting. Furthermore, the ultimate control over the centralized
resources and functions of the Internet wielded by the United States will likely be
challenged more frequently by those in the international community seeking truly
international governance of the Internets shared infrastructure.
The Internet community faces significant collective challenges in maintaining
and upgrading that shared infrastructure to meet ever-increasing user expectations.
An optimal level of coordinated planning of the Internets common infrastructure lies
somewhere between the complete lack of centralized control of content and the strong
centralized control of the DNS. The problems caused by the requirement that each
domain name be globally unique may eventually be overshadowed by the exhaustion
of the pool of IP addresses that IPv4 provides, unless IPv6 can be implemented in a
timely manner. However, as explained in the standards and protocols sublayer
discussion above, that implementation does not appear to be happening, at least in
North America. If the advantages of IPv6 are lost, then so are the operational and
associated regulatory advantages over traditional physical layer networks.
D. Physical Layer
The content, application, and operational layers all ride over the physical layer.
Physical layer elements, like content layer elements, tend to be subject to legal
regimes that predate the Internet. In other words, physical layer elements are
governed in the same way now as they were before the Internet joined them into one
interoperable, interconnected whole. In this section, some of these legal regimes and
616
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
their applicability to each of the equipment and networks sublayers will be identified.
The specific example of IP telephony is used to illustrate the types of considerations
relevant to legal analysis at the physical layer.
1. Equipment Sublayer
The Internet equipment industry can be thought of as a branch of the computer
industry, which predates the Internet. The great achievement of TCP/IP, of course, is
in linking different types of computers running different operating systems.
Therefore, equipment can be designed independent of upper-layer considerations.
Similarly, the computer law field continues to exist independent of Internet law,
thanks in part to the modularity of Internet architecture. For this reason, the field of
computer law, which is dominated by patent, copyright, and trade secret law, is not
considered further in this paper.
2. Networks Sublayer
independently of
The regulation of public telecommunications networks also predates the Internet
and continues
is being
progressively undermined.171 As the Internet and IP networks continue their apparent
convergence into a single, unified platform for public and private communications,
the physical layer will be influenced more and more by what is going on above it.
That is, higher-layer issues will likely force changes in physical layer technology and
policy, as is already the case in the broadband access realm.172
independence, however,
it. This
The incredible and largely unplanned success of the Internet has surprised
everybody, including the communications industry and its commentators. Dreams of
ubiquitous computing and communications abounded in the 1990s, but the visions
tended to be of centrally controlled systems run by traditional telephone and cable
companies. The buzzword was convergence, and the challenge was to determine how
telephone and cable companies could be carefully allowed to enter each others
markets.173 The theory was that both industries would convert to digital data networks
that could carry voice as easily as video, rendering distinctions between the two
arbitrary.174 Then along came the Internet, at first operated by universities and
171 The best general resource in this area is Sunny Handa et al., Communications Law in Canada,
looseleaf (Markham, Ont.: Butterworths, 2000).
172 See e.g. the discussion of IP telephony regulation in Part II.D.3, below.
173 The primary policy documents of the era were: CRTC, Competition and Culture on Canadas
Information Highway: Managing the Realities of Transition (19 May 1995); Industry Canada,
Building the Information Society: Moving Canada into the 21st Century (1996); Information Highway
Advisory Council (IHAC), Preparing Canada for a Digital World, Final Report of the Information
Highway Advisory Council (1997).
174 See e.g. Sheridan E. Scott & David B. Elder, Changing Communications Regulation in the
Information Age (2000) 14 Can. J. Admin. L. & Prac. 153; Sheridan Scott, Regulation on the
Information Highway: Capturing the Elusive Butterfly (1995) 9 Can. J. Admin. L. & Prac. 305; Alan
Ross & Jennifer Pawson, Crossing Guards on the Electronic Highway: The Basis for Federal
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
617
institutions, but later joined by commercial ISPs. The Internet offered easy access to
an unfathomably vast store of unadulterated and free information. While more
expensive and sophisticated systems offering video-on-demand and satellite
telephone service have come and gone, demand for the Internet continues to grow
(though not at boom era levels).175
The CRTC has been dealing with the Internet in explicit terms (that is, using the
term Internet) since at least 1996, being among the first regulators in the world to
do so. The CRTC now oversees an industry that has achieved the second highest
broadband Internet penetration in the world, behind South Korea and ahead of
Sweden.176 The CRTC considers at least the Internets content and physical layers to
be very much within its jurisdiction, but has generally opted not to regulate. In its
landmark New Media public notice of May 1999, the CRTC ruled that the
Broadcasting Act applies to some content available over the Internet, but not to all
content.177 The CRTC interpreted its broadcasting jurisdiction widely but agreed to
issue an exemption order,178 leaving content consisting only of audio, video, a
combination of audio and video, or other visual images including still images that do
not consist predominantly of alphanumeric text179 unconditionally (but not
permanently) unlicenced. Content that consists predominantly of alphanumeric text
(e.g., e-mail, most Web pages) remains outside the reach of the Broadcasting Act.
With respect to its Telecommunications Act jurisdiction, however, the CRTC has
not shied away from taking action to encourage the development of the Canadian
Internet service market. Initially, ISPs were thought of by the bellheads at the
telephone companies as parasites on the PSTN, benefiting from its ubiquity but not
contributing to it (in the form of universal service funding support). While this view
was not strictly accurate (as explained below), ISPs enjoy a privileged place in
Canadas telecommunications regulatory regime.180 The restrictions traditionally
placed on common carriers, such as non-discrimination among customers and traffic,
made the PSTN a relatively open platformone that could be used by unregulated
Jurisdiction Over Convergence Technology (1994) 3 Dal. J. Leg. Studies 209; Hudson N. Janisch,
Optical Fibre to the Home: Keeping Our Policy Options Open (1991) 3 Windsor Rev. Legal Soc.
Issues 1.
175 Statistics Canada, supra note 1.
176 OECD, Directorate for Science, Technology and Industry, The Development of Broadband
Access in OECD Countries DSTI/ICCP/TISP(2001)2/FINAL (Paris: OECD, 2001), Table 4
(Broadband status (June 2001)), online: OECD
177 Supra note 142. See also Michael Koch, CRTC Regulation of the Internet: Carriage Yes,
Content No in Osgoode Hall Law School, Information Technology and Cyberspace Law (Toronto:
Emond Montgomery, 1999) c. 23.
178 Exemption Order for New Media Broadcasting Undertakings (17 December 1999), Public
Notice CRTC 1999-197, online: CRTC
179 Broadcasting Act, supra note 141, s. 35.
180 Due in part, no doubt, to political aversion to killing the goose that laid the golden egg.
618
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
service providers largely as they pleased. Early commercial ISPs simply could order
large numbers of business telephone lines, hook modems up to them, advertise a dial-
in number, and they were in business. ISPs have directly benefited from these and
other favourable regulatory policies. More recently they have been granted certain
positive rights of access to the broadband Internet access facilities of both telephone
companies and cable television companies.
The CRTC has attempted to take a very aggressive approach to opening up
broadband network facilities to independent ISPs so that they can resell high-speed
Internet services and compete with ISPs that are vertically integrated with the owners
of the two dominant local access infrastructures in Canada (twisted copper pair and
coaxial cable lines). With respect to cable TV companies, the CRTC ruled in July
1998 that it would not regulate the rates at which broadcast carriers (cable companies)
offer retail level Internet services. It would, however, require companies to file tariffs
setting out standard terms on which they would provide competitive providers of
retail services with access to the cable companies telecommunications facilities.181
With respect to Digital Subscriber Line (DSL) service, independent ISPs can also
purchase, at discounted rates, those elements of telephone company high-speed
Internet facilities that they need to provide competing DSL services. The state of
competitiveness in the market for DSL services (and hence the need for CRTC
intervention), however, has always been a topic of contention. For example, the
independent members of CAIP182 filed an application with the CRTC in August, 2001
alleging that they were unable to enter the residential market for high-speed DSL
services in Ontario and Quebec due to anti-competitive conduct by Bell Canada and
its affiliates. While the commission refused to restrict the Bell companies pricing
flexibility, it did impose winback rules to prevent them from trying to lure back
former customers who have gone with a competitor for ninety days after the date of
disconnection.183 This type of restriction is typical in the local and long distance
telephone service markets and illustrates the degree to which the CRTC is willing to
apply traditional telecommunications regulatory tools and concepts to the Internet
market.
181 Broadcast Carriers, supra note 165; Regulation Under the Telecommunications Act of Cable
Carriers Access Services, (6 July 1999), Telecom Decision CRTC 99-8, online: CRTC
Large Cable Carriers Higher Speed Access Service (21 August 2000), Telecom Order CRTC 2000-
789, online: CRTC
Rates Approved for Large Cable Carriers Higher Speed Access Service (31 January 2001), Telecom
Order CRTC 2000-789-1, online: CRTC
182 That is to say, those that are not affiliated with the telephone or cable companies.
183 Independent Members of the Canadian Association of Internet ProvidersDigital Subscriber
Line Internet Services by Bell Canada and Bell Nexxia (27 June 2002), Telecom Decision CRTC
2002-37 at para. 63, online: CRTC
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
619
The most significant privilege that the Internet industry has enjoyed is its
exemption from the contribution regime that supports universal access to basic
telecommunications services. ISPs have always contributed implicitly by paying fees
for underlying telecommunications services, including elements of subsidy funding.
Retail Internet service itself is exempt. This is in stark contrast to the broad range of
other telecommunications services for which these charges must be paid.184 The place
of one particular Internet application within this contribution regime is the subject of
the fourth and final layer example.
3. Example: IP Telephony185
One of the most anticipated innovations touted by the netheads of the Internet
industry in the 1990s was IP telephony. While it has not yet taken off to the extent
expected (and certainly not in the way its narrowband cousin, e-mail, has), many in
the telecommunications industry believe that it is only a matter of time before all
telephone calls travel in IP form.186 IP telephony is a software application that can be
run on any packet-switched data network (including the Internet). It allows users to
simulate real-time telephone calls using various combinations of computers and/or
telephones.
There are generally two types of IP telephony: Internet telephony, where the
underlying, long-haul transmission network is the Internet; and Voice-over-IP,
where the underlying, long-haul transmission network is any kind of packet-switched
network other than the public Internet (often a private corporate network). Like e-
mail, IP telephony is also an application. Instead of being dependent on the
intervention of a service provider, however, IP telephony in most cases utilizes the
interface between the Internet and the telephone system. Thus, IP telephony is most
usefully analyzed in the context of the physical layer.
IP telephony combines audio digitization and compression technologies with
Internet technology. While traditional telephone systems relied on electrical impulses
to reproduce speech, modern telecommunication is based on the digital encoding of
sound into streams of bits. The bits travel over defined paths (or circuits) from
origin to destination. These paths are controlled by the interconnecting telephone
companies that jointly carry a given telephone call. For this reason, traditional
184 Changes to the Contribution Regime (30 November 2000), Decision CRTC 2000-745 at para.
91, online: CRTC
revenues from wholesale Internet services are subject to the contribution charge, the amount of lost
subsidy funds is not as significant as it might sound.
185 This section is derived in part from Craig McTaggart, IP Telephony and Canadas
Telecommunications Regulatory Regime (2001/2002) 2 Internet and E-Commerce Law in Canada
49 [McTaggart, IP Telephony].
186 Tiffany Kary, Net Telephony Poised to Take Off? CNet News (31 May 2002), online: CNet
News
620
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
telephony is referred to as circuit-switched, and interconnecting networks are
collectively referred to as the PSTN.
Packet-switched data networks, such as those running IP, carry streams of
digitized sound that have been chopped up into packets. The originating device (such
as a computer) sends the pieces of a telephone call over one or more packet-switched
data networks, and a computer on the other end reassembles them and plays back
the sound. When the process is repeated in the opposite direction, a live two-way
voice conversation (or fax communication) can be carried on. Successive generations
of IP telephony technology allowed this process to take place, first between two
computers (known as PC-to-PC), then between a computer and a telephone (PC-
to-Phone), and finally between two telephones. To facilitate such Phone-to-Phone
calls, computer servers, referred to as gateways, are employed and convert calls
from circuit-switched mode to packet-switched mode (and vice versa).
IP telephony provides a new way of transmitting phone and fax calls, a service
which has traditionally been highly regulated in most countries.187 In Canada, the
primary regulatory issue presented by IP telephony has been its impact on the explicit
subsidy regime that supports universal service. The system of internal cross-subsidies
and implicit subsidies that prevailed during the monopoly era was gradually made
explicit with the onset of competition in Canadian telecommunications markets. That
system, known as the contribution regime, has undergone several major changes in
the past decade, the details of which are not essential for present purposes.188 As a
result of several key decisions,189 Canada does not regulate IP telephony
specificallyin fact, most forms can be described as unregulated. Gross revenues
above a certain threshold, however,190 from some forms of IP telephony, depending on
how the calls interact with the PSTN, are subject to an annual contribution charge,
currently set at 1.1 percent.191
Whether the contribution charge is levied depends on where a given call is
converted from packet-switched to circuit-switched form. If that conversion takes
place within Canada, such that an ordinary telephone located within Canada is used at
187 For information on the international regulatory status of IP telephony, see Regulatory Aspects
of IP Telephony in International Telecommunication Union, ITU Internet Reports: IP Telephony
(Geneva: International Telecommunication Union, 2000) c. 4.
188 See McTaggart, IP Telephony, supra note 186.
189 Telecom Order CRTC 97-590 (1 May 1997), online: CRTC
Internet Service Providers (17 September 1998), Telecom Order CRTC 98-929, online: CRTC
supra note 185.
190 That is, where the service provider in question has ten million dollars or more of such revenues
in the year. See Changes to the Contribution Regime, ibid. at paras. 97-100. This effectively leaves
small IP telephony operators outside the contribution regime.
191 Final 2003 Revenue-Percent Charge and Related Matters (19 December 2003), Decision CRTC
2003-84, online: CRTC
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
621
one or both ends of an IP telephony call, any associated revenues earned in Canada
are subject to the contribution charge. Any person offering international long distance
services to the public, whether by means of IP telephony or otherwise, is required to
hold a licence and report their eligible contribution revenues.192
IP telephony has never been illegal in Canada, as it still is in many countries such
as Botswana, Nepal, and Turkey.193 It cannot be blocked by telephone companies,
even if they would prefer that their customers not use IP telephony, lest it reduce
circuit-switched long distance revenues. SaskTels high-speed Internet terms of
service contain the odd stipulation: The Service is not intended to be used for long
distance services.194 The nature of plain vanilla Internet access, however, is that
users can do whatever they like with that access simply by employing the applications
of their choice: the Internet empowers its users. In its decisions regarding the
regulatory status of IP telephony, the CRTC has walked a fine line between
preserving this freedom and ensuring that long-standing policy goals such as
universal service continue to be met. We can expect more of these kinds of policy
challenges as the Internet becomes more and more ingrained in Canadian society. The
degree to which those in control of elements at one layer can or should be allowed to
influence elements at other layers will likely be a frequent point of contention.
While higher layers can operate largely independently of the physical layer, they
are increasingly affected by physical layer regulation. The Internet began its popular
life as an alternative to, or even enemy of, the traditional telephone networks. Yet
gradually, and particularly when the speed of telephone lines became a barrier to
Internet growth, this independence was eroded. Now that the largest ISPs in Canada
(and most of the world) are also the largest telephone and cable TV companies,
Internet service is just another service regulated by the CRTC. The market is still the
primary guarantor of consumer interests in this area. As the importance of Internet
service to Canadians continues to grow, however, calls for greater CRTC involvement
may be expected, particularly as the distinctions between old and new forms of
communication continue to disappear.
IP telephony regulation is an interesting example of this transition as it is a case
of the physical layer regulator imposing terms on the operation of an application-layer
business. The CRTC explicitly recognized that it is either impossible or very difficult
to identify the contents of each packet to see whether it carries bits of a contribution-
eligible service. Instead, the regulation of IP telephony is focussed on the operation of
servers that convert traffic for passage back and forth between traditional circuit-
192 Regulatory Regime for the Provision of International Telecommunications Services (1 October
1998), Telecom Decision CRTC 98-17, online: CRTC
Contribution IssuesFinal Consensus Report (17 December 1999).
193 International Telecommunication Union, supra note 188 at 36.
194 SaskTel, SaskTel High Speed Internet Terms of Service, online: SaskTel
622
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
switched networks and Internet-style packet-switched networks. One can speculate
that future Internet-related issues that resemble telecommunications issues, such as
the operation of instant messaging networks, might also lead the CRTC to project its
physical layer jurisdiction onto application layer matters. On a related note, the layer
at which an element notionally resides can also significantly alter the question of
legislative jurisdiction. The United States ultimate control over the Internets
centralized resources and functions (which Canada has recently shifted from
recognizing to protesting mildly) can be contrasted with the control over physical
networks that national regulators like the CRTC continue to hold.
The case of independent ISP access to telephone and cable company broadband
facilities, on the other hand, suggests that a hands off policy, while popular with
respect to higher layers, may do more harm than good at lower layers. It is widely
believed that a cable and DSL duopoly would result from the removal of the
independent access rules, since incumbent network owners would have no incentive
to help independent ISPs compete. Worse, with a stranglehold on most users access
to the Internet (at least through fat pipes), the physical network operators would be
in a position to manipulate those elements of the operational layer that they also
control. The need for competition among cable and broadband service providers
illustrates how arguments and policy values relevant to one layer may be harmful if
applied indiscriminately to other layers.
Lawful access to ISP facilities is an example of another phenomenon: that of
addressing legal issues indirectly at other layers. Law enforcement agencies are
seeking to interpose a measure of control at the ISP functions sublayer and at the
physical layer to counteract the freedom from regulation provided by the application
layer. Given the right degree of access to the underlying data streams, the contents of
ones Internet sessions can be captured and studied. Users, however, may respond (if
they are aware of such scrutiny) by employing encryption tools at the application
layer. The only sure way to guarantee the security of ones computer, both against
spying from the outside in the case of lawful access and from the inside in the case of
surreptitious applications, is to unplug it from the Internet. This is an option whose
appeal and even feasibility seems to be continually diminishing.
Conclusion
The reality of constant connection to the Internet will continue to bring to the fore
both familiar and novel legal and policy issues in the Internet context. The layered
conceptual model put forward in this article is intended to provide a useful way of
thinking about the Internet in that context. The article is premised on the idea that
when confronting such issues, decision-makers, policy-makers, and legal analysts
should identify the precise issue before them and consider its possible links to, and
implications for, other elements of the Internet. A view of the Internet that includes
the content, application, operational, and physical layers (and their respective
sublayers) supports a much more comprehensive understanding of Internet legal and
policy issues.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
623
There are at least three ways in which this four-layer approach to Internet legal
and policy analysis may prove useful in future work. First, identifying the specific
elements of Internet architecture implicated in any given issue may provide insights
into preliminary analytical considerations such as identifying parties, applicable legal
regimes, and relevant jurisdictions. Second, since the architectural and legal
circumstances of the layers can differ markedly, one must be careful not to let the
characteristics or prevailing wisdom regarding one layer unduly influence ones
analysis of elements at other layers. Third, since control of elements at one layer can
have consequences for elements at other layers, it is important to understand these
elements on their own terms, and in particular, to identify who controls them.
A layered approach to the Internet can help in the preliminary identification of
parties, applicable legal regimes, and relevant jurisdictions by drawing attention to
the many different elements of Internet architecture that may be implicated by any
given issue. For example, the most common Internet application, e-mail, requires
software applications on users computers to work in conjunction with software on
their ISPs servers, in the case of hosted e-mail accounts, or with those of a Web-
based service provider, in the case of Webmail. These servers can theoretically be
located anywhere on the Internet. In turn, both forms of e-mail service depend on
the availability of key centralized resources and functions such as the DNS and the
RSS. Assigning blame for service problems is complicated by the fact that so many
different actors are involved, the legal rights and obligations of whom generally
remain vague in this relatively new domain.
This article has sought to broaden the readers appreciation for the number of
different elements of Internet architecture (and therefore independent actors) that
must function together to make the Internet work. In many regards, we are still in the
early stages of sorting out the legal status of, and relationships among, these elements
and those who operate them. Fundamental questions of what legal regimes apply or
should apply remain to be explored. Ironically, though perhaps not surprisingly,
national jurisdiction over physical layer matters is being seized upon as a means of
indirectly addressing higher-layer issues, such as IP telephony. This interplay between
layers and legal regimes highlights the degree to which the architectural and legal
circumstances of the layers can differ, making it important to keep the characteristics
and considerations relating to the various layers and sublayers separate during analysis.
In particular, it is important to bear in mind the contrast between the bountiful
diversity of content and transactions at the top layer, on one hand, and the technical
and economic constraints typical of the operational and physical layers, where
uniformity or scarcity are more often the dominant features. While the Internet can
theoretically support an unlimited number of Web sites about books, for example, it is
generally thought to be capable of supporting only one of each domain name.
Similarly, the number of different applications that users can employ on the Internet is
limited only by their creative ability. Yet the number of options that the average user
624
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 48
has for connecting to the Internet is usually limited, particularly outside of urban areas.
Diversity and choice characterize the Internets upper layers, but not its lower layers.
A particularly interesting avenue for future research would be a consideration of
the implications of control of elements at one layer for elements at other layers. For
example, the roles of the providers of operational layer infrastructure, such as the
centralized resources and functions and ISP networks, need exploration. The effects
of the extension of power over the physical layer into the application and content
realms are potentially troubling. On the theory that who pays the piper calls the
tune, end-user control over the application layer, for example, should be jealously
guarded against the encroachment of television-style systems where all programming
choices are in the hands of one-way broadcasters. Of course, some consumers can be
expected to demand simpler versions of the Internet, such as America Online, being
content to give up some degree of edge-based control in favour of centralized
programming. As suggested by the discussion of overlay software, the demands of
advertisers sometimes clash with the interests of individual users. This conflict has a
long history in other media, and is often characterized by surreptitious invasions of
unwitting users privacy. Control of Internet elements is an important issue, given the
number of different elements involved and the highly dispersed nature of that control.
The extremely wide range of legal and policy issues canvassed in this article
themselves only a fraction of existing Internet-related problemsillustrates the
tremendous breadth of the field. This breadth is all the more remarkable given that the
Internet entered the public consciousness less than ten years ago. Whether Internet
law exists as a separate field of law or merely brings together related issues from
many existing fields, continues to be a subject of debate.195 Without taking a position
on that question, which perhaps only time can answer, this article argues that a
catholic approach to Internet legal and policy analysis should be preferred over a
blinkered one, given the impact of the Internets layered architecture on the analysis
of any given issue.
195 See Frank H. Easterbrook, Cyberspace and the Law of the Horse [1996] U. Chi. Legal F. 207;
Lawrence Lessig, The Law of the Horse: What Cyberlaw Might Teach (1999) 113 Harv. L. Rev.
501.
2003]
C. MCTAGGART INTERNET LEGAL ANALYSIS
625
ARPANET
Advanced Research Projects Agency Network
BGP
CAIP
CRTC
DNS
DSL
EDI
gTLD
HTTP
IANA
ICANN
IETF
IHAC
IP
ISP
NIC
Border Gateway Protocol
Canadian Association of Internet Providers
Canadian Radio-television and Telecommunications Commission
Domain Name System
Digital Subscriber Line
Electronic Data Interchange
generic Top-Level Domain
Hypertext Transfer Protocol
Internet Assigned Numbers Authority
Internet Corporation for Assigned Names and Numbers
Internet Engineering Task Force
Information Highway Advisory Council
Internet Protocol
Internet Service Provider
Network Interface Card
NSFNET
National Science Foundation Network
PC
PSTN
RBUA
RSS
SMTP
SOCAN
SWIFT
TCP/IP
UDRP
Personal Computer
Public Switched Telephone Network
Residential Broadband Users Association
Root Server System
Simple Mail Transfer Protocol
Society of Composers, Authors, and Music Pulishers of Canada
Society for Interbank Financial Telecommunication
Transmission Control Protocol / Internet Protocol
Uniform Domain Name Dispute Resolution Policy
