might become providers of web content (as they presently may not).1 2These ISPs will have the
power to discriminate in the choice of Internet services they allow, and customers who want
broadband access will have to accept their choice. Giving this power to discriminate to the owner
of the actual network wires is fundamentally inconsistent with End-to-End design.
This technological bundling at the network level puts pressure on the principle of
End-to-End. These cable-owned-ISPs would thereby influence the development and use of cable
broadband technology. They would be exercising that influence not at the “ends” of the network,
but at the center. They are therefore shifting control over innovation, as Dr. Jerome Saltzer has
written, from a variety of users and programmers to a single network owner. This design defeats
the principle that the network remain neutral, and empower the users. It is the first step to a
return to the architecture of the old AT&T monopoly.
The Costs of Violating Architectural Principles
The costs of violating this fundamental principle of the Internet’s design are hard
to reckon. We simply do not know enough to know how sensitive the innovation of the Internet
is to changes in this competitive architecture. Obviously, in part the significance turns on how
the broadband market develops. But given trends as we can identify them now, the risks of
significant consequences from this design are great. We detail some of these risks below.
The first is the cost of losing ISP competition. As we have argued, one should not
think of ISPs as providing a fixed and immutable set of services. Right now ISPs typically
provide customer support, as well as an IP address that channels the customer’s data.
Competition among ISPs focuses on access speed, as well as some competition for content.
12These limitations are imposed by At Home Corporation. See @Home Acceptable Use Policy.
http://www.home.com/support/aup/(Visited November 8, 1999); At Home Corporation. @Home Frequently Asked
AOL, for example, is both an access provider and content provider. Mindspring, on the other
hand, simply provides access.
In the future, however, ISPs are potential vertical competitors to access providers
who could provide competitive packages of content, or differently optimized caching servers, or
different mixes of customer support, or advanced Internet services. This ISP competition would
provide a constant pressure on access providers to optimize access.
The benefits from this competition in the history of the Internet so far should not
be underestimated. The ISP market is extraordinarily competitive. This competition has driven
providers to expand capacity and lower prices. It has also driven providers to give highly
effective customer support. This extraordinary build-out of capacity has not been incented
through the promise of monopoly protection. The competitive market has provided a sufficient
incentive, and the market has responded.
The second cost is the risk that legacy business models will improperly affect the
architecture of the net. Broadband is a potential competitor to traditional cable video services.
Traditional cable providers might well view this competition as a long term threat to their
business model, and they may not want to change to face that competitive threat. By gaining
control over the network architecture, however, cable providers are in a position to affect the
development of the architecture so as to minimize the threat of broadband to their own video
market. For example, a broadband cable provider that has control over the ISPs its customers use
might be expected to restrict customers’ access to streaming video from competitive content
sources, in order to preserve its market of traditional cable video.
The third cost of such control by a strategic actor is the threat to innovation.
Innovators are less likely to invest in a market where a powerful actor has the power to behave
strategically against it. Innovation in streaming technologies, for example, is less likely when a
strategic actor can affect the selection of streaming technologies, against new, and competitive
One example of this cost to innovation is the uncertainty that is created for future
applications of broadband technology. One specific set of such applications are those that count
on the Internet being “always on.” Applications are being developed, for example, that would
allow the net to monitor home security, or the health of an at-risk resident. These applications
would depend upon constant Internet access.
Whether, as a software designer, it makes sense to develop such applications
depends in part upon the likelihood that they could be deployed in broadband cable contexts.
Under the End-to-End design of the Internet, this would not be a question. The network would
carry everything; the choice about use would be made by the user. But under the design proposed
by the merged company, AT&T affiliates would have the power to decide whether these
particular services would be “permitted” on the cable broadband network. Cable has already
exercised this power to discriminate against some services. They have given no guarantee of
non-discrimination in the future. Thus if cable decided that such services would not be permitted,
the return to an innovator would be reduced by the proportion of the residential broadband
market controlled by cable.
Our point is not that cable would necessarily discriminate against such
technologies. Rather, our point is that the possibility of such discrimination increases the risk an
innovator faces when deciding whether to design for the net. The increasing risk is a cost to
innovation, and this cost should be expected to reduce innovation.
Perhaps some of these costs could be remedied by competition from other
broadband providers. If cable companies restrict the nature of ISP service for broadband cable,
then to the extent there is competition from DSL, DSL might have a competitive advantage over
But not all of the costs to the Internet market from this change in architectural
design could be remedied by competition from other broadband providers. In particular, the cost
to innovation would not be remedied by competition among providers. That cost is borne by the
market as a whole, not by particular consumers in the market. Consumers individually don’t feel
any cost from this threat to innovation. They therefore have no additional incentive to move from
one kind of provider (cable) to another (DSL). Thus, if the increase in strategic power dampens
the willingness to invest in broadband technologies, there is no mechanism by which that effect
will be felt, and remedied, by broadband consumers directly.
More importantly, given the approach that the FCC has adopted for this case in
particular, there is no reason to expect that the cost will be avoided in other cases. As each new
broadband technology enters the Internet market, the FCC’s position in this case would imply
that that new technology too could violate this principle of End-to-End design. Only xDSL
would be required (because of existing statutory obligations) to maintain the principle of End-to-
End design with respect to ISP choice.1 3And even if xDSL does provide a competitive market
for some ISPs who want to serve broadband access (on which more below), it simply makes no
sense as a matter of economic policy to foreclose the largest possible market for ISP competition,
particularly when doing so serves no good end.
The Importance of Acting Now
As we describe more fully below, there are those within the FCC who have
expressed the view that there is no reason for the FCC to address the open access question in the
context of this merger. The merger itself will not change the bundling policy of the existing
AT&T Cable Services network. Thus any problem with open access, some would claim, is not
exacerbated by the merger.
This view misunderstands the potential for strategic action. If there are five
broadband cable networks, each acting independently, then the threat to innovation is less than if
these five broadband cable networks could act in unison. If they were independent, then the
decision of some networks to block certain kinds of Internet services would not necessarily
influence any other networks. Thus the threat to innovation would not be as great. Once the cable
monopolies can act together, however, and decision to discriminate would affect a larger section
of the market. The risk to innovation would therefore be much greater. Further, AT&T is
implementing its bundling policy now, and a firm stance in favor of open access by the FCC
could have a beneficial effect on AT&T’s policy, not only regarding MediaOne, but in other
The “wait and see” approach also discounts the cost of regulating ex post. In its
present state, the ISPs that AT&T would rely upon are independent business units. If the merger
were completed, they could easily be folded into the resulting entity. Once integrated, the
13Further, if bundling of broadband service is permitted for every network except those based on classic telephone
wires, eventually xDSL providers will have a strong moral case that they should not be subject to a restriction that
regulatory costs of identifying non-discriminatory rates would be much higher than they would
be under the existing structure. Rather than the complexity that DSL regulation involves,
imposing a rule of open access now would be relatively less costly. The same is even more true
of independent ISPs. If the vibrant market for ISPs in narrowband access is weakened or
destroyed because they cannot provide broadband service, those ISPs and their innovative
contributions will disappear. If they do, we won’t magically get competition back by deciding
later to open the broadband market to competition.
A Comparison to United States v. Microsoft1 4
To see the significance of the threat in the context of broadband, it is useful to
compare the nature of the bundle at issue in this merger with the threat that the government has
alleged in United States v. Microsoftthat Microsoft poses. Obviously the two cases are different
in many ways. Microsoft’s operating system is far more dominant than is cable broadband
service. But the point of the comparison is not to equate the competitive threat of the two
services. It is to see the structural equivalence between the threats.
The government’s primary claim against Microsoft is a charge of monopoly
maintenance. The argument is that Microsoft bundled its browser with its operating system, so as
to foreclose effective competition in the browser market, and thereby protect its monopoly
returns in the operating system market. The threat that the government claims Microsoft was
avoiding was the development of a robust application platform, built around Java technologies.
As the browser was the platform within which such applications could develop, it was important,
the government argues, to keep control of the browser market.
does not burden any of their competitors.
The issues in United States v. Microsoftare extremely complex. No fair
consideration of the real issues in the case could conclude that either side has an easy argument.
But what is clear is that the behavior alleged against Microsoft is far less controlling than the tie
In this matter too, the claim of those supporting open access is that
AT&T/MediaOne would be in a position to maintain monopoly power, at least over the video
market. Like the Microsoft case, this maintenance would be affected by keeping control over the
source of potential competition. In the Microsoft case, that was the browser; in this case, that is
But importantly, the level of control exercised by AT&T in this case is far greater
than the control Microsoft is alleged to assert. The government has never argued that Microsoft
totally disabled the ability of competing browsers to be installed on client machines; the most the
government alleged was that Microsoft made it difficult, or uneconomical, to load a competing
browser. Once properly installed, a competitor browser on the Windows platform works just as
well as Microsoft’s. Or more precisely, the government has not alleged that the platform disables
In the case of cable broadband, however, the architecture does disable the relevant
competition. One simply cannot choose a competitor ISP as the primary ISP in the cable
broadband architecture, and thus one cannot choose a competitor to provide the primary ISP
14 We note that Lessig served as special master in a prior proceeding between the United States and Microsoft, and
Lemley has served as a consultant to the Antitrust Division of the United States Department of Justice on the current
case. It is not our attention to offer here any opinion on the merits of either case.
AT&T argues that this competition is not disabled by the cable broadband
architecture, since a customer can always “click-through” to a non-cable ISP. But the ability to
click through provides just a fraction of the services that a competitor ISP might potentially
provide. It would be as if competitor browsers on the Windows platform performed just 30% of
the functions that they performed on other platforms. Further, click-though may be economically
irrational even if it is technically feasible, just as Microsoft’s original “per processor” license
made it nominally possible but extremely unlikely for an OEM to load two operating systems
onto a computer. Thus the question in this matter is not whether a user will take the time to
“download” another ISP connection; there’s no such download possible. The architecture ties the
user to AT&T/MediaOne’s ISP; users cannot cut that knot.
The Arguments in Favor of Broadband Bundling
A Policy of “Regulatory Restraint”
It is our view that AT&T’s desired design of the architecture of the emerging
broadband cable market could be a significant threat to innovation in this market. We suggest a
presumption that no significant portion of the broadband market be permitted to violate the
“End-to-End” design, unless there is clear evidence that such a change is benign.
So far the FCC has taken a different view. In its initial consideration of this
matter, and in the most recent reports from the Cable Services Bureau, the FCC has taken the
position that it would best facilitate competition in this market by simply doing nothing. In our
view, this profoundly underplays the importance of the FCC’s activism in assuring competition
in the past, and will jeopardize the innovative prospects for broadband Internet service in the
The Cable Services Bureau most recent report to Chairman William Kennard has
recommend a policy of “regulatory restraint.”15It grounds its recommendations on a number of
“responses and preliminary findings,” and on a straightforward cost benefit analysis of the risks,
and benefits, from “regulatory restraint.” The “responses and preliminary findings” are as
The broadband industry is nascent;
Cable modem deployment spurs alternative broadband
Regulation or the threat of regulation ultimately slows
deployment of broadband
Market forces will compel cable companies to negotiate
access agreements with unaffiliated ISPs, preventing cable
companies from keeping systems closed and proprietary
If market forces fail and cable becomes the dominant
means of Internet access, regulation might then be
necessary to promote competition
There was no consensus on how to implement “open
access” from a regulatory perspective
There was no consensus on how to implement “open
access” from a technical perspective
Rapid nationwide broadband deployment depends on a
In our view, conclusions (1) and (2) are correct. Conclusions (6), (7) and (8) may
be correct, but are irrelevant to this proceeding. Conclusions (3), (4), and (5), the heart of the
policy recommendation, are both wrong and internally inconsistent.
Findings (1) and (2): It is clearly correct that broadband services are just
beginning. The vast majority of Internet users are narrowband users. The content and services
15See Broadband Today, Staff Report To William E. Kennard, Chairman Federal Communications Commission On
Industry Monitoring Sessions Convened By Cable Services Bureau (October 1999).
that fit best with broadband are just being developed. These services are in part services that
require large bandwidth to function effectively. (Streaming video or audio is an example). More
significantly, they are also services that assume that the user is “always on” the Internet. This
latter fact will, in our view, lead to the most significant change in how the Internet will be used.
There are a host of applications that are just beginning to be envisioned that will depend upon the
Internet constantly monitoring and responding to situations “at home.” Many of these services
are difficult or impossible to implement through modem-based telephone access.
It also appears correct, though we have not studied the matter independently, that
cable broadband service has spurred other broadband providers, in particular DSL. We do
believe the report overstates the significance of existing DSL competition. The current market
share of cable in the residential broadband market is over 80%.1 6This lead is significant, and is
unlikely to change quickly.1 7
Findings 6 and 7: The Bureau maintains that there is neither agreement on how to
implement “open access” nor agreement on what “open access” is. But this part of the report
reads like a poor imitation of a Socratic dialogue. Obviously, if one gathers a collection of bright
lawyers and technologists, each advancing different interests, one can create a cacophony of
views about what “open access” is, just as a good law professor can create a cacophony of views
about what “justice” is, or even what the “FCC” is. But a law professor can not deny that there is
an “FCC” merely because no “agreement” in definition is found. And the Bureau should not
17See, e.g., Forrester Report, From Dial-Up to Broadband, April 1999, at 10.