ARCHIVED — National Association of Broadcasters

COPYRIGHT REFORM PROCESS

SUBMISSIONS RECEIVED REGARDING THE CONSULTATION PAPERS

---

Documents received have been posted in the official language in which they were submitted. All are posted as received by the departments, however all address information has been removed.

---

Submission from the National Association of Broadcasters received on September 17, 2001 via e-mail

Subject: Comments of the National Association of Broadcasters

PDF Version


September 17, 2001


Comments - Government of Canada Copyright Reform
c/o Intellectual Property Policy Directorate
Industry Canada
235 Queen Street
5th Floor West
Ottawa, ON K1A 0H5

By E-mail: copyright-droitdauteur@ic.gc.ca

The National Association of Broadcasters (NAB), a non-profit incorporated association consisting of commercial television and radio stations and broadcast networks which serves and represents the United States broadcasting industry, is pleased to provide its comments with respect to the Government of Canada's Consultation Paper on the Application of the Copyright Act's Compulsory Retransmssion License to the Internet.

Respectfully submitted,



Henry J. Baumann
Executive Vice President / Law & Regulatory Affairs

Benjamin F.P. Ivins
Senior Associate General Counsel

Enclosures

I. INTRODUCTION.............................................................................. 1

A. National Association of Broadcasters............................................................................................. 1

B. Summary of Position....................................................................................................................... 1

II. OVERVIEW OF U.S. BROADCASTING
SYSTEM............................................................................................. 2

III. WHY INTERNET RETRANSMISSION
OF T.V. STATION SIGNALS IS SO
DANGEROUS............................................................................................. 4

IV. CLAIMS THAT TECHNOLOGICAL
MEANS CAN EFFECTIVELY LIMIT
INTERNET RETRANSMISSION OF BROADCAST
SIGNALS TO SUBSCRIBERS TRULY RESIDING
WITHIN CANADA ARE UNFOUNDED............................................................................................ 7

A. Access Control Systems Are Inevitably Prone to Error............................................................................... 8

B. Access Control Systems Will Inevitably Be Bypassed............................................................................................. 9

1. Proxy Servers............................................................................................. 10

2. Remote Access Systems............................................................................................. 10

3. Deep Linking............................................................................................. 10


V. A CANADIAN COMPULSORY LICENSE THAT
ALLOWS PARTIES OUTSIDE CANADA TO
RECEIVE U.S. BROADCAST SIGNALS WOULD
VIOLATE INTERNATIONAL LAW............................................................................................. 12

VI. CONSENSUAL INTERNET TRANSMISSIONS OF
RADIO BROADCAST SIGNALS ARE ENTIRELY
DIFFERENT AND DISTINGUISHABLE FROM
NON-CONSENSUAL RETRANSMISSION OF
TELEVISION BROADCAST SIGNALS AND DO
NOT JUSTIFY A COMPULSORY INTERNET
RETRANSMISSION LICENSE FOR TELEVISION
SIGNALS............................................................................................. 14

VII. TECHNOLOGICAL NEUTRALITY CANNOT
JUSTIFY A LICENSE TO RETRANSMIT
TELEVISION SIGNALS VIA THE INTERNET............................................................................................. 16

VIII. SIGNAL "LEAKAGE" BY CABLE AND
SATELLITE RETRANSMITTERS IS NOT
COMPARABLE TO, AND DOES NOT JUSTIFY,
INTERNET BROADCAST RETRANSMISSIONS........................................................................... 18

IX. A "QUALIFIED TERRITORIAL RESTRICTION"
WILL NOT REMEDY THE FATAL FLAWS IN
AN INTERNET RETRANSMITTER COMPULSORY
LICENSE............................................................................................. 20

A. Flaws In Reasonable Technological Measures
To Effectuate Territorial Restrictions............................................................................................. 20

B. Flaws In Monitoring Requirements To
Effectuate Territorial Restrictions............................................................................................. 21

X. THE PROPOSALS TO REQUIRE ONLY
"EFFECTIVE CORRECTIVE ACTION" ARE
UNWORKABLE AND INEQUITABLE............................................................................................. 23

XI. THE FULL PANOPLY OF COPYRIGHT
INFRINGEMENT REMEDIES FOR VIOLATION
OF ANY TERRITORIAL RESTRICTIONS IS
ESSENTIAL............................................................................................. 24

XII. BANNER ADVERTISING SHOULD NOT
BE PERMITTED IN ANY EVENT............................................................................................. 28

XIII. CONCLUSION............................................................................................. 29

APPENDIX A

APPENDIX B

Application of the Canadian Copyright Act's Compulsory Retransmission License to the Internet

1. Introduction
   
   
   1. National Association of Broadcasters

      The NAB is a non-profit incorporated association of television and radio stations and broadcast networks which serves and represents the United States broadcasting industry. NAB includes among its members over 1,100 U.S. television stations.

   2. Summary of Position

      The NAB commends the Copyright Policy Branch, Canadian Heritage and the Intellectual Property and Policy Directorate, Industry Canada (hereafter "Departments") for their thoughtful and comprehensive review of the issues and summary of arguments in the Consultation Paper On The Application Of The Copyright Act's Compulsory License To The Internet (hereafter "Consultation Paper").

   Succinctly stated, the NAB is unalterably opposed to the non-consensual retransmission of television broadcast signals over the Internet by means of a statutory or compulsory license. Internet retransmissions of United States broadcast signals by Canadian entities pursuant to Section 31 of the Canadian Copyright Act could result in devastating consequences for the United States system of free over-the-air broadcasting and could violate the Berne Convention to which both Canada and the United States are signatories. The Departments are totally and completely correct in their conclusion that:

   "non-consensual retransmission to locations outside Canada of the protected works contained in retransmitted television signals does not further Canadian public policy interests. To the contrary, it risks harm to the legitimate interests of rights holders who would be obliged to bring proceedings in multiple jurisdictions in the hope of preserving the integrity of their territory-specific licensing arrangements. Although it has been argued that recourse to foreign law ought to be the solution for rights holders concerned with the preservation of their territory-specific "business models", the departments believe this could be an inequitable consequence of a compulsory licensing regime."

   While the Departments go on to conclude that "if Internet-based retransmissions were to have the benefit of the License it likely must, at a minimum, be subject to an appropriate territorial restriction", NAB will demonstrate that currently there are no technological measures to assure that any territorial restrictions, even were they imposed on paper, could be enforced. Indeed, the essential nature of the Internet medium -- that it is global, interconnected, and non-hierarchical -- works directly against any attempts to implement geography-based restrictions, and it was this marked difference between the Internet and other media that led the United States to decline to extend statutory licenses to cover Internet retransmissions. Accordingly, NAB urges that Canadian law be clarified to expressly exclude from its retransmission license the retransmission of television broadcast signals by Internet retransmitters.

   
   
2. Overview of U.S. Broadcasting System

   The success and viability of the U.S. television broadcasting system exists as a result of the partnership between national networks, program syndicators and local television stations. Under this system, local TV stations in markets large and small provide a combination of national TV programming, syndicated programs, and local news, weather and public affairs programming.

   To describe this system more fully, to explain the crucial importance of local program exclusivity to its continued viability, and to analyze how streaming of programming on distant television signals into local markets would undermine the system, NAB retained Dr. Mark Fratrik, Vice President of BIA Financial Network. Dr. Fratrik's analysis is attached at Appendix A.

   As Dr. Fratrik explains, the continued vitality of this system depends on local stations enjoying a substantial degree of exclusivity in providing network and syndicated programming to local viewers. Local stations make most of their revenues by selling advertising time during popular network and syndicated programs. During these same programs, local stations also run promotional spots designed to attract viewers to local news programs. These spots are a key way that stations build audiences for their news programs.

   Protection of stations from importation of duplicative programming into their markets is thoroughly woven into the fabric of our legal system. Since the 1960s, for example, the Federal Communications Commission has adopted and enforced network nonduplication, syndicated exclusivity, and sports blackout rules that bar cable systems from importing duplicative programming from distant stations. Congress acknowledged and incorporated these rules in 1976 as conditions for the cable compulsory license , and reaffirmed its strong support of these rules in the Telecommunications Act of 1996.

   When satellite television appeared on the scene, Congress created a similar set of rules in 1988 to protect the network/affiliate relationship. Congress reaffirmed those rules in 1999 in the Satellite Home Viewer Improvement Act, and directed the Commission to apply syndicated exclusivity and sports blackout rules to satellite carriers as well.

   
   
3. Why Internet Retransmission of T.V. Station Signals Is So Dangerous

   As Dr. Fratrik explains, allowing third party Internet services to retransmit U.S. television station signals over the Internet could cause incalculable harm to those stations, and put their basic economics in grave jeopardy. The reason is simple. If local viewers are able to watch network and syndicated programs on distant stations imported over the Internet, the value of local stations being the exclusive provider of that programming in their market will be lost. Advertisers will pay much less to sponsor programming on a local station, if that programming is also available day and night on other U.S. and foreign television stations being imported over the Internet from all over the world. Moreover, time zone differentiations will mean that viewers in western time zones could view programming being streamed from a Buffalo, New York station at an earlier time and never even tune into their local stations.

   The inevitable reduction in advertising revenues resulting from the streaming of distant signal programming into local markets will have multiple negative consequences. Stations will be forced to cut back on their spending, for both nationally syndicated programming and their local programming, including local news and informational programs. This, in turn, will have a negative effect on program suppliers, as they too adjust and reduce the amount and quality of programming they supply. Cuts could also be expected in discretionary local station spending on community service and outreach programs.

   The ultimate real loser in this downward spiral will be viewers, particularly those who cannot or choose not to subscribe to pay multichannel video services.

   The dimensions of this issue are huge. In the United States alone, there are 110 million Americans with Internet access, and there are hundreds of millions of Internet users worldwide. These figures are, of course, growing every day, as is access to broadband that makes delivery of video over the Internet even more appealing to consumers. Local stations would face Internet companies delivering broadcast programs -- the same programming for which they have licensed local exclusive rights -- to viewers in their local markets, including many of the most affluent households that are the most appealing to advertisers.

   And it is not just the national programming that is at issue here. For example, Buffalo, New York stations invest millions of dollars every year in producing top quality local news programming, of which they are the sole copyright owner. Having invested the money and hired the talent to produce this programming, these stations, and not some third party who has invested not a penny in their programming, are entitled to decide how to deliver that programming to viewers.

   Nor is this just a domestic issue. Delivering television programming throughout the world under an uncontrolled compulsory license on the Internet, transmission of TV stations would sabotage the ability of Canadian and U.S. broadcasters and other copyright owners to sell their programming in foreign markets. That is, all owners of television programming would find themselves "scooped" in selling their own programming by third parties who could simply appropriate the entire output of television programming and deliver it instantaneously through the world.

   It is not an exaggeration to say that Internet transmissions of TV stations could cripple, if not destroy, the U.S. and Canadian system of free, local, over-the-air television. Exclusivity would become a meaningless concept. Dozens, or even hundreds, of stations could be streamed over the Internet offering the same programming that local stations once used as their calling card.

   The damage that can be caused by retransmission of television broadcast signals over the Internet under a compulsory licensing regime was succinctly described by the U.S. Register of Copyrights in recent testimony before the U.S. Congress as follows:

   "Because of the ease with which copyrighted materials can move around the world on the Internet, the defeat of a keylock system can spell instant disaster for these works. Broadcast programming intended for limited television markets within the U.S. could become available worldwide with no control or compensation for the copyright owner. Further, even if protection devices are in place to limit receipt of a broadcast signal from a source to a specific geographic location, there may be no control over the receiver of that signal that prevents him from further retransmitting the signal to others. Again, technological solutions may be developed to address these concerns, but until they are, and unless we can be confident of their reliability and security, enactment of a compulsory license for local signals would place broadcast programming in jeopardy."

   
   
4. Claims That Technological Means Can Effectively Limit Internet Retransmission Of Broadcast Signals To Subscribers Truly Residing Within Canada Are Unfounded

   In the Consultation Paper the Departments correctly and appropriately oppose a compulsory retransmitter license that would permit unrestricted retransmissions of broadcast signals over the Internet, but appear enticed by the notion that such a license might be acceptable if it were subject to "an appropriate territorial restriction." Such restrictions presuppose, of course, viable technological mechanisms to implement them. To assist the Departments in their review and analysis of this crucial practical issue, NAB retained Benjamin Edelman, who is a staff member of Harvard University's Berkman Center for Internet and Society. The Berkman Center is a nonprofit entrepreneurial research program at Harvard Law School, which was founded in 1997 to explore and understand cyberspace, share in its study, and help pioneer its development. We asked Mr. Edelman, who is expert in web server administration and is experienced in, among other areas, automated systems that attempt to determi ne geographic locations of Internet users, to identify and describe the limitations and challenges to a system purporting to restrict access to Internet retransmissions of broadcast signals to end users located within Canada. The results of Mr. Edelman's analysis are contained in Appendix B attached hereto. For the reasons discussed below, Mr. Edelman concludes that "if over-the-air television programming is retransmitted over the Internet, users located in the United States will be able to gain access to the retransmitted programming."

   The likely failure of technical measures attempting to restrict access by non-Canadian users stems both from inherent deficiencies in geographic analysis tools that impede their accurate detection of user locations and from the ready availability of multiple ways to circumvent such access controls altogether. In a way, the most essential characteristics of the Internet medium -- its global connectivity and its use of a system of multipath packetized transmission -- are fundamentally at odds with attempts to limit access to particular content within a single geographic area. Neither the structure nor the technology of the Internet supports geographic access restrictions, and the practical impediments described in more detail by Mr. Edelman reflect that fundamental fact. Indeed, as discussed below, it was this very incompatibility with geographical access controls that led the U.S. Copyright Office and Congress to conclude that the statutory retransmission licenses would not be extended in the United States to cover Internet retransmissions.

   1. Access Control Systems Are Inevitably Prone to Error

      As Mr. Edelman explains, the problems begin with the absence of any established location-based directory of the IP addresses attached to Internet devices. Access control companies must seek to infer locations from their own databases or from indirect information such as the path by which the user arrives at the site or the locations of network elements that appear to be associated with the user's IP address. For the reasons described by Mr. Edelman, these methods of inferring location from other information are bound to be unreliable. Importantly, this unreliability will be at its greatest when attempting to distinguish between Canadian and American users, because of the special nature and extent of Canada/U.S. operations by multinational companies and ISPs.

      There is no information available to confirm the accuracy of the inferences drawn by currently available geographic analysis tools, either in general or in the especially difficult context of making Canada/U.S. distinctions, but the methods they use are virtually certain not to produce completely accurate results. And the incentive of the retransmitter to avoid too high a rate of "false negatives" (i.e., denials of service to actual Canadian users because of ambiguous or uncertain geographic analysis results) will tend to work strongly against applying a standard that would reject all but the most certain Canadian users in order to assure that there are no "false positives" (non-Canadian users who are mistakenly given access).

   2. Access Control Systems Will Inevitably Be Bypassed

      But even if currently available geographic analysis tools were not so flawed, access would still be available to those who could circumvent the geographic control system itself. As Mr. Edelman describes, such means of circumvention -- developed for other, wholly legitimate purposes and readily available to many users today -- are becoming even more widely available.

      1. Proxy Servers

         The first means of circumvention, proxy servers, act essentially as an intermediary between an Internet user and the websites he or she is accessing. The effect is that the location of the user appears to be at the location of the proxy server itself. While a geographic analysis tool may undertake measures to identify and reject queries coming through proxy servers, there are practical and technical impediments to its doing so with complete accuracy. And as concerns about Internet privacy continue to mount, there will be an increasing incidence of users who choose to conceal their true identity and whereabouts as they visit various websites of all types.

      2. Remote Access Systems

         A second similar circumvention issue arises from the increasing use of remote access systems for corporate and other networks. "Tunneling methods" of access would routinely allow a user accessing the Internet through a company network to appear to be located at the geographic location of the tunneling server. "Terminal services methods" allow the remote operation of a computer for many purposes, but the effect would again be to present the geographical location of the computer being remotely operated, not the true location of the user.

      3. Deep Linking

         Even more significant, however, is the fact that geographical analysis tools are designed to work through the HTTP-based web server of the retransmitter, while the retransmitted programming itself is provided through separate streaming media server systems. This raises the possibility that users may, by various means, link directly to the streaming media content. As Mr. Edelman explains, the nature of the Internet is such that only a handful of such "deep linking" incursions by knowledgeable individuals could result in much more widespread access. In effect, a pinhole leak in the streaming video pipeline created by a handful of determined individuals would quickly and inevitably become a pipeline blowout as many thousands of users clicked on an easy link to access the programming. Based on Mr. Edelman's investigation and analysis, it does not appear possible at the present time to cut off this simple method for bypassing the retransmitter's geo graphic location access control system altogether.

   Mr. Edelman's conclusion that, presently, the retransmission of broadcast signals over the Internet by a Canadian retransmitter could not effectively be limited to Canada is consistent with the conclusion of the U.S. Register of Copyrights, as reflected in testimony before Congress:

   Some firms are working on software and hardware that would restrict the distribution of information, which could include broadcast retransmission, to specific Internet customers or to customers located in a specific geographic area. But no one has yet rolled out a fail-proof system, and if experience has taught anything with technological controls to copyright, it is that it is not long before they are hacked or circumvented.

   Even JumpTV implicitly admits there is no proven system to contain Internet retransmitted broadcast signals within Canada. In its application to the Copyright Board, JumpTV concedes that geographical blocking systems for the Internet "are being developed", and that it cannot assure during this "development stage", "that there would be no 'leakage' to some viewers located outside of Canada."

   Given the enormity of what is at stake, NAB respectfully submits that the U.S. and Canadian broadcasting industries should not be permitted to serve as experimental fodder for what are basically untested developmental geographic blocking mechanisms. It seems likely that determined users will find ways to bypass border control authenticator systems at relatively minimal cost, and that access will rapidly become possible via any number of techniques, as described in Mr. Edelman's paper.

   
   
5. A Canadian Compulsory License That Allows Parties Outside Canada To Receive U.S. Broadcast Signals Would Violate International Law

   Both the United States and Canada are signatories to the Berne Convention For The Protection Of Literary and Artistic Works (Berne) and the WTO Agreement on Trade-Related Aspects of International Property Rights (TRIPS). Among the obligations to which Canada is subject under Berne is the requirement that authors be granted the exclusive right to authorize "communication to the public by wire or by rebroadcasting of the broadcast of the work, when this communication is made by an organization other than the original one."

   While Berne authorizes countries to place conditions on the exercise of this exclusive right by imposition of a compulsory license, Berne also imposes three express limitations on the conditions countries can impose on such licenses: they "shall apply only in the country where they have been prescribed; [t]hey shall not in any circumstances be prejudicial to the moral rights of the author, nor to his right to obtain equitable remuneration."

   The reason for Berne's limiting the authorization provided by a country's compulsory license to its own borders was explained by the U.S. Register of Copyrights as follows:

   "This Berne limitation on compulsory license flows from the territorial nature of copyright laws. National copyright laws govern conduct within their respective territories. Authors and other copyright owners must look to the national laws of the country in which protection is claimed to determine their rights. A necessary corollary to this principle of territoriality is that the national copyright law of one country cannot authorize the exercise of exclusive rights in another country."

   The conclusion reached by the Register is correct and compelling: "Berne requires member countries to impose territorial limitations on retransmissions that are carried out under a compulsory license . . . In order to comply with Berne's territorial limitations on compulsory licenses, a [U.S.] compulsory license for retransmission of broadcast television signals on the Internet could only permit such transmissions for reception within the United States." Moreover, the Register has found that:

   "[A]s far as we have been able to ascertain through our discussions with the industry there is no technology at the present time that is one hundred percent effective at preventing reception of signals outside the boundaries of a particular country. Given the present state of the technology, it appears unlikely that we could implement a Berne-compatible compulsory licensing regime that permits unencrypted retransmissions of television signals over the Internet."

   These views and findings by the U.S. Register of Copyrights provide strong support for the proposition that authorizing the retransmission, under Section 31, of broadcast signals on the Internet so as to permit reception outside of Canada would violate Berne. And as established above, there exists no current means to restrict such signals to guarantee that they will not be received outside Canada.

   
   
6. Consensual Internet Transmissions of Radio Broadcast Signals Are Entirely Different And Distinguishable From Non-Consensual Retransmission Of Television Broadcast Signals And Do Not Justify A Compulsory Internet Retransmission License For Television Signals

   Section 2.6 of the Consultation Paper correctly and appropriately distinguishes between the "transmission" over the Internet by many radio stations of their own signals or the retransmission of such signals by third parties based upon consensual agreements with radio stations, and the non-consensual retransmission by third parties of television signals by means of a compulsory license at issue here. The first key distinction from the broadcaster's perspective is, of course, the consensual nature of the radio transmission as opposed to the proposed compulsory nature of the retransmission of television stations.

   Equally fundamentally, however, are the huge legal distinctions and historical differences between the way radio stations, and the music and sound recordings most of them perform, are licensed and marketed, and the way television stations and audiovisual works are marketed. As emphasized in the Fratrik Paper, a key element in the economics of a television station is the local exclusivity it obtains from most of its programming. It is of great importance and value to the advertisers of a program to know that multiple sources will not be importing that same programming into the market. This value to advertisers is also reflected in the compensation local television stations provide to their program suppliers by means of marketplace negotiations, not compulsory licensing.

   Radio, on the other hand, has a completely different history and licensing structure. No radio station, or radio station advertisers, expects the station to have the exclusive rights to any musical work even with respect to other stations in the market, much less with respect to distant market stations, hence Internet intrusion of other distant stations into the market playing the same music is of less significance. Moreover, the producers of music and sound recordings are accustomed to, and have built their business models around, licensing their product to stations under a compulsory license regime.

   It is important to understand these crucial distinctions in order to disabuse any notion that somehow because radio stations engage in Internet transmissions of their own signals, television stations should have no objections to having their signals involuntarily retransmitted by others over the Internet.

   
   
7. Technological Neutrality Cannot Justify A License To Retransmit Television Signals via the Internet

   The statutory copyright licenses are intended to balance the interests of the public in having copyrighted content made available to them with the interests and incentives of the creators of those copyrighted works. When the balance is right, works will be made available to consumers and the copyright owners will be able to receive sufficient compensation from both statutory licensees and other licensees with whom the owner can continue to negotiate in the marketplace. But the balance needs to be made with full regard for the nature of the compulsory licensee and the consequences of the grant. A principle of technological neutrality would not be neutral at all if it were to require that the same license be offered to radically different technologies. A license allowing non-consensual retransmission of television broadcast signals via the Internet would sow the very seeds of destruction of the copyright owner's interests and incentives, to the ultimate detriment of the consumers it was i ntended to favor.

   Under U.S. copyright law, the remarkably specific form of the various statutory licenses reflects the fact that each was a carefully crafted legislative compromise between the interests of copyright owners and those of the particular users or retransmitters, in light of the nature of the technology being used. For example, as the Copyright Office observed about the cable license, which was adopted after more than ten years of active consideration by Congress, "[t]he cable license is clearly the product of a very difficult legislative compromise, and the stripes of the compromise can be seen in the excruciating detail of section 111 of the Copyright Act."

   The availability of each of the statutory licenses is expressly contingent on the satisfaction of specified conditions, which are intended to permit the authorized retransmissions to be made without destroying the free broadcasting system on which they rely. The cable license, for example, is strictly limited to carriage of signals "permissible under the rules, regulations, or authorizations of the Federal Communications Commission." . In United Video, Inc. v. FCC, 890 F.2d 1173, 1184 (D.C. Cir. 1989), the U.S. Court of Appeals for the D.C. Circuit rejected cable operators' arguments that certain FCC rules prohibiting retransmission of syndicated programs for which broadcasters had negotiated geographically exclusive contractual licenses infringed the cable operators' rights under their Copyright Act statutory license. The court found that the cable operators "will not be able to obtain a compulsory license" to carry those programs in the first place, given the explic it Section 111 condition of compliance with FCC rules. Those rules were designed to allow the then-new cable technology to grow without destructive consequences for the system of free local broadcasting. In enacting an expressly contingent cable compulsory license, Congress "specifically decreed that the copyright liability of cable companies would be determined in part by the regulations of the Commission, created in furtherance of communications policy goals." These policy goals squarely support the system of free broadcasting in geographically exclusive local markets, upon which the system of geographical copyright exclusivity also is grounded.

   By contrast, the later-enacted satellite compulsory license in Section 119 of the U.S. Copyright Act imposed a detailed list of different conditions that addressed the very different nature of the satellite technology. These include a set of very detailed restrictions and procedures for implementing prohibitions on the retransmission of network signals to subscribers who could receive an affiliate of the same network over the air. These prohibitions serve the fundamental purpose of permitting satellite retransmissions of television programming only where doing so would not compete with a local network affiliate's own programming, to the detriment of the system of free over-the-air broadcast stations serving their local markets. In addition to these detailed prohibitions in the satellite compulsory license itself, Section 119 includes a provision making the license contingent on compliance with the FCC's rules. Thus, the scope and effect of the compulsory licenses are specifically tai lored to the differences in the two transmission technologies and to the different communications policy treatment accorded to the two transmission services by the FCC.

   
   
8. Signal "Leakage" By Cable And Satellite Retransmitters Is Not Comparable To, And Does Not Justify, Internet Broadcast Retransmissions

   The Departments observe that Canadian DTH satellite services are sometimes illicitly received outside of Canada, despite ongoing countermeasures. While this may be true, it certainly does not, as the Departments suggest, justify that Internet retransmitters also be permitted some vague and unspecified "free leakage." In her report to Congress, the U.S. Register of Copyrights identified the distinctions between "closed path retransmissions services" such as cable and satellite, for which the damage from such illegal and unauthorized "leakage" is limited and controllable, and illegal and unauthorized "leakage" from Internet retransmissions that can make the broadcast signal "suddenly available to anyone with a television set."

   Some have asserted that signal theft is signal theft, and that such activities on the Internet would be no different than signal theft of cable or satellite service. We disagree. Cable is a closed path retransmission service with little customer activity. While a consumer can obtain an illegal "black box" for cable, that consumer cannot use his or her equipment to defeat the encryption system for cable signals, nor can that consumer further retransmit the cable signals around the world. The same is true for satellite. One can obtain an illegal smart card to make a digital satellite box operate, but that equipment cannot make a satellite suddenly available to anyone with a television set. These activities can occur, however, on the Internet, which makes comparisons between retransmissions via cable or satellite and the Internet inapposite.

   Mr. Edelman's analysis further confirms that once retransmitted broadcast signals get on the Internet, there is simply no way to maintain or assure any minimal or acceptable level of unauthorized leakage. This phenomenon derives in part from the very nature of the Internet itself. The Internet is an interconnected global network of networks, designed to allow "packetized" transmissions to arrive at their destination through multiple paths. It is a medium of widely dispersed and diverse facilities, designed to work without a hierarchical network structure that would offer the opportunity for control. The Internet has also developed culturally as an environment of freedom and anonymity. Given the incentive to view desirable television programming and the technical means made possible by the Internet's unique technology and architecture, Internet users in the United States will surely be able to access the retransmitted programming.

   
   
9. A "Qualified Territorial Restriction" Will Not Remedy The Fatal Flaws In An Internet Retransmitter Compulsory License

   There is simply no merit to the suggestion that an Internet retransmitter license could somehow be made palatable and acceptable by imposing territorial restrictions based on the maintenance of "reasonable technological measures" and monitoring. The most fatal flaw in these proposals is that they presuppose technology to restrict reception to locations in Canada and methods to monitor the efficacy of such technology which, as Mr. Edelman has amply demonstrated, do not exist. It would be extraordinarily irresponsible to establish a legal mechanism that essentially invites and encourages parties to retransmit broadcast signals over the Internet based upon compliance with conditions they are almost certainly incapable of meeting.

   1. Flaws In Reasonable Technological Measures To Effectuate Territorial Restrictions

      The "reasonable technological measures" standard proposes to require transmitters to establish technological measures "which persons without specialist knowledge could not reasonably be expected to circumvent except with outside assistance." One of the many flaws with this standard is that many circumvention methods described by Mr. Edelman may only require a specialist to intervene at the onset to establish the method, which any persons without special knowledge could then exploit to circumvent. For example, the direct bypass of geographic access control systems by "deep linking" to the streamed content need be performed only once per channel, and the whole world's Internet users, whether novice or sophisticated, would be able soon to follow, without being required to perform or even understand the technical means of gaining access. This was done with respect to the iCraveTV site, and will happen again.

      A second flaw with this standard is that in this era of such rapid technological change, what requires "specialist knowledge" one day can become common knowledge the next. As more and more Internet users become concerned with protecting their privacy in Internet commerce, for example, more and more will use proxy server sites to be able to control the disclosure of their identities, and will thus become familiar with one of the techniques that may also thwart a geographic access control system for television programming.

      A third flaw with this approach is that it will lead to interminable legal wrangling and litigation about such issues as what qualifies as "specialist knowledge" or "intelligible reception."

   2. Flaws In Monitoring Requirements To Effectuate Territorial Restrictions

      The proposed monitoring requirement suffers from similar flaws. Chief among these is that which the Departments themselves acknowledge, namely that "illicit reception is largely hidden reception which is only partially addressed by monitoring obligations and audit privileges." Moreover, it is not clear that the geographic location analysis tools are even able to identify the extent to which their systems fail properly to identify Canadian-based end users.

      Another flaw with the monitoring proposal is its suggestion that only Canadian retransmission collectives be responsible for monitoring compliance with technical restrictions. In the U.S. context, most of the victims of unacceptable leakage, and those that would suffer most, are not members of these collectives, but rather local U.S. television stations into whose markets the programming will be leaked. Under the proposal, are all of these stations supposed to audit the monitoring results of each and every Canadian retransmitter (assuming the monitoring results would even provide intelligible information about their markets)? And if a retransmitter exceeds some established level of leakage of, for example, a CBS signal to Internet users in Glendive, Montana, what is the CBS affiliate KXGN in Glendive supposed to do? Commence a lawsuit or administrative proceeding in Canada? In Glendive? The reason Glendive was chosen as an example is it is the smallest DMA in the Unit ed States. But there are hundreds of stations in small and middle sized markets - those stations that could be most economically devastated by "leaking" - for whom monitoring and enforcement would be wholly impractical.

      
      
10. The Proposals To Require Only "Effective Corrective Action" Are Unworkable And Inequitable

    The Consultation Paper's proposals for "effective corrective action," both "particular" and "systemic," are unworkable and inequitable. The proposal that a retransmitter would have to take corrective action against "particular instances of foreign reception" (however that would be defined ) only if it could be expected, given the nature of its operation, to have the means of addressing individual violations, would enable a retransmitter to easily cloak itself in total immunity from enforcement obligations. All a retransmitter would have to do to avoid any such obligation is to design its system -- which is supported by advertising and hence has an incentive to maximize user access, not restrict it -- so that the system is incapable of dealing with individual violations.

    As for the proposed required corrective actions for systemic flaws in a retransmitter's anticircumvention measures, the Departments themselves point out the insurmountable problem with establishing any violation "threshold." More fundamentally, however, the suggestion that some "commercial scale" threshold be established below which access by users in the U.S. would be permitted is simply unprecedented and would do violence to the Departments' stated principle of providing an "equitable balance among stakeholders [by][ ensuring that rights holders are treated in a fair and equitable manner." What is at issue here are copyright ownership interests and, specifically, a proposal to give a unique new class of commercial users of copyrighted material the already exceptional benefit of a non-consensual, compulsory, statutory license. Now, what is being proposed on top of that is some pre-defined acceptable level of cheating -- some actual amount of acknowledged copyright infringement -- for which the benefiting retransmitters will incur no copyright liability whatsoever. The granting of a compulsory license with knowledge that it will lead to infringement and injury to U.S. broadcasters and the U.S. system of local broadcasting should hardly be contemplated, let alone implemented.

    
    
11. The Full Panoply Of Copyright Infringement Remedies For Violation Of Any Territorial Restrictions Is Essential

    The proposal to permit Internet retransmissions on the basis of territorial protections fully acknowledged to be ineffectual would be unjustifiable. But the proposed truncation of legal remedies that would be available to try to staunch the tide of purposefully permitted infringements would be unconscionable. NAB is totally opposed to "the view that non-compliance with a territorial restriction should not render the entire retransmission an infringement of copyright, subject to the full range of existing remedies for direct infringement, but rather available remedies must be related specifically to non-compliant foreign reception." Moreover, in response to the Departments' specific question, limiting remedies to injunctive relief would be utterly inadequate and would constitute an open invitation to massive abuse. There is no reason or justification why the full panoply of remedies for copyright infringement, including attorney's fees and statutory damages, should not be ava ilable to parties aggrieved by violations of any territorial restrictions, and such remedies should be available both in Canada and in the jurisdiction(s) of reception. It is also imperative that summary procedures be adopted to allow aggrieved parties to obtain expeditious and effective relief.

    As NAB has demonstrated, there currently are no technological mechanisms to implement territorial restrictions effectively. Yet, if Canada were to expand its retransmitter license to allow for the Internet retransmission of broadcast signals, it would serve as an open invitation for all manner of enterprises to use Canada as a test bed for untried and untested technologies. Scores of such entities could appear on the scene, with no capitalization requirements, leaving them essentially judgment proof. To the extent these entities would use advertising to support their service, not only would there be absolutely no incentive to invest in effective territorial restriction technology or to monitor its effectiveness, but the incentives would be just the opposite, namely to allow as much leakage as the retransmitter thinks it could get away with. The burden of monitoring and enforcement would be left to rights holders who would be left with the extraordinary burden and expense of tracking dow n and suing these entities. As demonstrated by Dr. Fratrik, the damages such companies would inflict on broadcasters could be swift and enormous. One can foresee that they would simply shrug their shoulders at the activities of proxy servers, deep link sites and others over whom they would claim no control. And it would then fall to aggrieved rights holders to track all these contributory infringers down and sue them individually. Limiting infringement remedies to injunctions would be unconscionable. No sooner would an injunction be obtained against one of these entities than it would simply shut down and another would open up somewhere else.

    Given the magnitude of damage that could be wreaked upon rights holders by widespread violation of any territorial restrictions, the suggestion that the remedy of an injunction requiring a retransmitter to terminate all retransmissions until it is capable of complying with territorial restrictions might be too severe is stunning. NAB is at a loss to understand how the Departments could deem such a remedy to be inconsistent with Canada broadcasting law and policy, when six pages earlier in the Consultation Paper they correctly found that "non-consensual transmission to locations outside Canada of protected works contained in retransmitted television signals does not further Canada public policy interests." Many of the damages and harms described by Dr. Fratrik to which U.S. broadcasters would be subject hold equally true for Canadian broadcast interests, an outcome which, indisputably, would be "inconsistent with Canadian broadcasting law and policy."

    NAB finds equally stunning and disconcerting the Departments' apparent conclusion that even if a retransmitter became aware that one of its authorized recipients within Canada was using the Internet to retransmit and redirect a broadcast signal outside of Canada, presumably in blatant disregard for any territorial restriction, the retransmitter should not be required by law immediately to discontinue service to such an authorized recipient. The Departments' explanation for this inexplicable conclusion is that the discontinuance of service to such a scofflaw: "could have the disproportionate effect of denying Canadian households access for broadcasting purposes."

    First, NAB is at a total loss to understand how imposing such a minimal, and fundamentally fair, requirement as telling retransmitters that if they discover a subscriber breaking the law (violating territorial restrictions), they must remove the means by which they are doing so, "disproportionately" deprives Canadian households access to broadcasting services. The message to retransmitters and households should be clear and simple: violate the territorial restriction and you lose the privilege of both sending and received Internet retransmitted signals.

    Second, absolving retransmitters of any responsibility for the malfeasance of the recipients of their signal, provides a massive opportunity to make a mockery of any territorial restrictions that were imposed. Given the economic incentives of advertiser supported retransmitters to ignore any geographic restrictions in the first place, compounded by the relative ease by which Canadian recipients could, in fact, redirect retransmissions outside of Canada, makes this proposal a recipe for anarchy.

    Not only should retransmitters be required immediately to discontinue service to any recipient the retransmitter knows, or has a basis for determining, has illegally redirected the signals, they should be required to report the names of such miscreants to copyrights holders and authorities for prosecution.

    
    
12. Banner Advertising Should Not Be Permitted in Any Event

    The harms that would result if non-consensual retransmission of television broadcast signals over the Internet were permitted would be even more greatly exacerbated if retransmitters were at the same time permitted to employ banner advertising. At the very least, express prohibitions against such advertising would be necessary in any retransmitter statutory license.

    The gross inequity of requiring a broadcaster to compete against a retransmitter for advertising revenues when the retransmitter is providing nothing more than the broadcaster's own programming is self evident. This inequity is particularly egregious when one considers that broadcasters must charge advertising rates high enough to recoup their substantial programming and other costs, while a retransmitter could charge much lower rates in light of its virtually non-existent costs.

    Similarly obvious is the untenable unfairness of a situation in which the advertising, for example, of car maker X appeared in a broadcaster's program while a competitive car maker's ad was appearing in a banner ad that framed the broadcaster's program on the Internet user's screen. Indeed, one could foresee a retransmitter using such an opportunity as a selling point.

    Then there is the inevitable confusion that would result to the public over who was sponsoring what. For example, U.S. broadcasters have a sponsorship identification requirement that requires them to announce when material is broadcast for money, services, or other valuable consideration and by whom or on whose behalf such consideration was supplied. The obvious statutory reason for the requirement is to let the public know that they are being solicited and by whom. With banner advertising, one could easily foresee instances where the identification of the sponsor of broadcast material would be muddled and confusing.

    Finally, allowing the use of banner advertising could result in a violation of the Berne Convention. Article 12 of Berne provides authors with "the exclusive right of authorizing adaptation, arrangements and other alterations of their work." This so-called "right of adaptation" is an exclusive right not subject to limitation through compulsory licensing. Banner advertising could result in the creation of an unauthorized adaptation under Article 12 of Berne.

    
    
13. Conclusion

    Permitting the retransmission of television broadcast signals over the Internet through a national compulsory license would violate the principles established by the Departments governing their deliberations in this proceeding, would have profound international implications and adverse consequences to both Canadian and non-Canadian copyright owners and broadcasters, and likely would violate international law.

    Creating a compulsory license that would permit the retransmission of television broadcast signals would violate all four of the governing principles enunciated by the Departments in that:

    • such a license would endanger any assurance that all Canadians now have of continuing to receive diverse cultural content provided by broadcasters because it would undermine the local market program exclusivity that is essential for broadcasters to continue to provide such content;
    • the proposals in the Consultation Paper for territorial restrictions, corrective actions, remedies, banner advertising, and unauthorized retransmission by authorized users are totally ineffectual and would result in gross unfairness and inequitable treatment both to rights holders and broadcasters;
    • a retransmitter compulsory license treating the Internet as if it were the same as cable and satellite would not be "technologically neutral" at all, it would extraordinarily disadvantage and crippled "old" broadcasting technology which still serves as the primary source of news and entertainment to millions of Canadians and Americans.
    • such a license would create great uncertainty and chaos to the current regime of broadcast program licensing that has provided a system of free over the air broadcasting to Canadians and Americans that is second to none. Moreover, the proposals for territorial restrictions, corrective actions, remedies, banner advertising and permitting so-called "authorized users" to engaged in unauthorized retransmissions would create a morass of legal, regulatory and jurisdictional ambiguities behind which unlawful activity could easily be sheltered.

    U.S. broadcasters have spent years and millions of dollars in litigation against its domestic satellite industry enforcing laws and regulations designed to protect local stations from incursions by satellite delivered distant signals violating local stations' rights to program exclusivity in their markets. The damage done by our domestic satellite industry pales by comparison to the havoc that could be wreaked upon our system of free over-the-air television in the U.S. were Canada to permit the retransmission of U.S. stations over the Internet.

    Accordingly, NAB urges the Departments in the strongest possible terms to adopt the position that Canadian law be clarified expressly to exclude from its retransmission license the retransmission of television broadcast signals over the Internet.

Respectfully submitted,


Henry L. Baumann
Benjamin F.P. Ivins
NATIONAL ASSOCIATION OF
BROADCASTERS
1771 N Street, N.W.
Washington, D.C. 20036


September 17, 2001






APPENDIX A


Mark R. Fratrik, Ph.D.
Vice President
BIA Financial Network
September 17, 2001



The Impact on U.S. Television Broadcasters
From Third Party Streaming of Local Television Stations

Introduction
The system that has evolved over the nearly sixty years of American commercial television is one of hundreds of local television markets whose local audiences are being served by local broadcasters. While many of these stations are affiliated with national networks and receive programming from their parent network and other providers of nationally syndicated product, these local stations supplement that programming with locally originated programming (e.g., local news programming) and are ultimately responsible for what is broadcast over the local station.
By having these affiliations with national networks and syndicators, the local television stations establish a local identity and brand affiliation, a major challenge in an ever increasingly competitive video marketplace. The local exclusive market areas for this programming provides an economic foundation for the local station to invest in its own programming and serve its local community.
The potential impact of altering that local brand identity with a third party streaming those signals via the Internet is very dramatic. Losing audiences to the streamed distant signals would directly lead to losses of advertising revenues, the only source of revenues for local television broadcasters. The breakdown in the exclusive brand identity will also severely weaken the local broadcaster.
Together these impacts could send the local television station into a downward spiral of losing revenues leading to less money available for programming investment, which in turn would lead to further losses in audiences and revenues. Another important secondary effect would be on the national program suppliers themselves, as they may receive less from local broadcasters. Consequently, they too would invest less in development.
The purpose of this paper is to investigate the potential effects resulting from the third party streaming of distant television stations on U.S. local television stations themselves. We first explain the existing marketplace for local U.S. television broadcasters – how they generate revenues and how they secure and produce programming. Next, we exmaine the impact of losing audiences from importation of distant signals into local stations’ markets by Internet streaming companies. Primarily this impact is from losing audiences that they can sell to advertisers, but the long term impact involves their potential investments in their own programming, and subsequent advertising revenues. These actions by local television stations impact the ultimate programmer developers, whether local or national, as they too may end up spending less. We conclude with some overall predictions as to what may occur if this third party streaming is permitted.
I. EXISTING U.S. LOCAL TELEVISION MARKETPLACE
Local television stations and exclusive market areas
There are presently 1,325 full-power commercial and 373 full-power non-commercial television stations in 210 television markets and other areas serving the U.S. population. As part of their license obligations, these television stations are required to serve these local communities through their programming and other outreach activities. The U.S. Congress has recognized the importance of these stations reaching their viewers in these markets by enacting laws for carriage obligations on satellite and local cable systems to carry all local television signals. Cable systems are required to offer carriage to all stations in the markets they serve, while satellite carriers must offer carriage to all local station in markets where they offer carriage to any station in that market.
An important part of the rationale for these carriage obligations as well as other FCC regulations protecting the exclusive program rights of local television stations is the role played by these stations to a significant part of the population. The latest estimate for the percentage of homes not subscribing to any multichannel provider is 20.9%. The percentage of television receivers not hooked up to these multichannel services is 30.3%. Statistical Research Inc., Home Technology Ownership Report, Spring 2001. In particular, the informational and entertainment services are essential to those who are unable and/or unwilling to subscribe to some pay multichannel video service providers. Without strong local free-over-the-air television stations, the specter emerges of a societal divide between those who have access to quality information and entertainment programming and those who do not. For the latter group, local television stations are lifelines.
In almost all of these markets, local television stations have exclusive rights to some nationally and regionally supplied programming. There is generally only one station affiliated with each of the national networks in each of these markets.. There are a few rare cases where there are two separate companies with affiliations with one national network, but that is due to the wide geographic area of the specific market. These affiliates determine whether to air (clear) all or some of the programming provided by the parent network. Additionally, there is generally only one station in each market that has a contract for specific syndicated programming, whether it is a package of movies or sports, previously-aired network originated programming, or programming airing for the first time.
This network and syndicated programming is especially important for local stations, as they tend to be the most popular and audience-attracting programming, thereby generating a substantial portion of the station’s revenues. Additionally, local television stations count on those audiences watching that programming also to watch station promotional messages of other programming and to stay on the channel for programming immediately following the syndicated and network programming, particularly the station’s local news.
These market exclusive arrangements are common among many other industries and have long been viewed as a pro-competitive arrangement. Retail outlets of major manufacturers (e.g., automobile tires) are routinely promised in contracts that they may have an exclusive area in which to sell. Because they know that they have this exclusive area, local outlets, including broadcasters, will benefit from locally promoting that brand and investing in their local operations. As a result, there is more competition on the local level between competitors, and consumers end up being the winner.
Exclusive market areas promote competition and are pro-consumer when that practice strengthens the final seller of the product (in this case local television stations) and there are many final sellers (providers of video programming). This pro-competitive outcome is certainly the case with local television stations. Excusive market areas for the network and syndicated programming establish these local television stations by providing a local “brand name” in the increasingly competitive video marketplace. Having the ability to be the “XXX Network’s” affiliate or the local station carrying certain syndicated programming provides the sound footing that stations build upon with their own local programming. These television stations are stronger when they face the competition from other local television stations and the hundreds of video choices now being provided by cable and satellite services.
The U.S. Congress and FCC have long seen the benefit to consumers from strong local broadcasters by instituting laws and regulations to protect the local station’s program exclusivity against dilution resulting from importation by cable and satellite services of the same programming into a local station’s market. Most recently the U.S. Congress in 1999, in the passage of the Satellite Home Viewer Improvement Act, reiterated its view, first expressed in 1988, on the importance of protecting these exclusive programming arrangements from the importation of satellite delivery of distant stations carrying network programming. Similarly, the FCC, for many years, has had syndicated exclusivity (Syndex) and network non-duplication rules that prevent local cable systems from importing out-of-market television signal programming into markets of local broadcast stations that have acquired exclusive local rights to that programming. These regulations are explicitly established to guarantee that local television stations have exclusive market areas for the different types of programming supplied by national providers (networks and national syndicators).
Local Television Station Revenues, Expenses, and Profits
In addition to the market exclusive network and nationally syndicated programming, local broadcast stations also invest considerable sums on their own local programming. Foremost among that commitment to investing in local programming is the commitment to providing local news. In 1999, the average television station spent more than a fifth (21.9%) of their operating expenses on local news operations. This result is from the annual survey of all commercial television stations in the U.S. See 2000 NAB/BCFM Television Financial Report, National Association of Broadcasters, Washington, D.C., p. vi. In some cases, that percentage can be as much as a third or more. Ibid., p. 3. Local programming investments do not stop at local news programming, but also in programming of other local interest.
With these investments in local programming as well as the investment and commitment from national program suppliers, local television stations are in the business of selling mass audiences to national and local advertisers. The ability to deliver a mass audience to potential advertisers is the comparative advantage that local television stations have in the advertising marketplace. That advertising marketplace has, however, become more competitive in recent years with the introduction of new competitors, e.g., new national and local cable channels and Internet sites. While many local television stations still are able to provide that mass audience product, the differences between the audiences they provide and audiences provided by other mass media are becoming smaller. Hence, the comparative advantage presently held by local television stations is in peril if the erosion of audiences continues or is exacerbated by actions of other industries.
This increased competitive marketplace has put severe strains on the profitability and viability of local television stations. One point must always be considered when examining and evaluating the profitability of local television station - the tremendous fixed cost nature of the operations that is independent of the size of audience it attracts. Because of these high fixed costs, any reductions in revenues lead to much more significant reductions in the profitability of local television stations.
For example, the costs of actually transmitting the over-the-air signal by and large are not dependent upon the size of the market in which it operates. The costs of the transmitter and well as the electricity to operate that transmitter will be the same if there are one million households in the market or if there are only one hundred thousand households.
Other costs at a television station are also fixed in the short term and are also independent of the final audience that is attracted. These costs include the contracted programming costs of the local television stations for specific syndicated programming. Of course, the syndicators and owners of this programming will charge higher prices to television stations located in larger markets with larger potential audiences. But, once those contracts are signed, any underperformance in delivering expected audiences do not result in lower programming costs for the station, only lower revenues and lower profits. In many cases, local television stations include guaranteed minimum audiences to advertisers for particular programming. If those guarantees are not met, local stations must provide additional minutes of advertising time, referred to as “make-goods,” that will thus be unable to be sold to other advertisers. Often these contracts are for multiple years, typically for three years.
The strain on local television station profitability was very noticeable even prior to the recent downturn in the American economy and advertising marketplace, especially in the smaller markets. For example, over half of all stations in markets ranked 71 through 80, and one quarter of all stations in markets 61 and smaller (through market size 210) had pre-tax losses in 1999. See 2000 NAB/BCFM Television Financial Report, National Association of Broadcasters, Washington, D.C., pp. 14-33. This state of affairs has only become worse as local television stations are under obligations to transition to digital facilities with expensive upfront costs while the economy has sputtered and the advertising marketplace has seen a substantial decline. These costs for digital transition are exceptionally burdensome for the mid-sized and smaller market stations that start out with a lower potential revenue base.
Impact Of Losing Audiences to Third Party Streaming
Smaller audiences and the impact on station revenues
As mentioned above, the comparative advantage enjoyed by local television stations is the ability to sell documented mass audiences to local and national advertisers. That comparative advantage has been shrinking in recent years from the onslaught of new choices available to consumers. Consequently, any further reductions in those audiences caused by distant television stations being streamed on the Internet could have a significant impact on the ability of local television station to capitalize on that comparative advantage and compete in today’s advertising marketplace.
Audience diversion may be especially significant if a local broadcast station from another market, possibly one from another time zone, is streamed. If consumers have the ability to time shift their viewing by watching a network program at another time, then substantial losses of viewers to all of the programming of the local broadcast station, including its own local programming, may occur. This programming is typically the most attractive and audience attracting programming, generating a substantial portion of the station’s revenues. Losing even a small portion of those audiences will have a considerable negative impact on the station’s revenues. The decrease in a station’s revenues will be immediate with the station having to provide additional inventory for “make-goods,” and long term as the station may have lower audience ratings numbers to sell future advertising spots.
In addition, the local broadcast station also counts on the audiences watching the last hour of prime-time network programming to continue watching that same channel leading into their own local late-night news programming. If viewers were watching that identical programming via a third party Internet stream, they would not necessarily watch the local news programming. This smaller audience for their local news programming will also greatly affect the station’s revenues.
Moreover, the importation of that other local broadcast station network and syndicated programming would tear down the exclusive market areas established by the program suppliers and local television stations. Broadcasters could no longer promote that they are the exclusive provider in a particular market of a certain type of programming, e.g., a particular network. What is also significant about this distant station being streamed into other markets is that while it is harming the local stations, the imported station may not benefit from that importation. It seems that Nielsen Media Research, the only television audience ratings company will not be able to quantify and report accurately those streamed signals. Moreover, even if there was accurate reporting of those audiences, the imported station will not benefit as the advertising marketplace is generally bought and sold on a local market basis. In other words, a local retailer would not care and certainly would not pay extra for the audiences in some distant market watching his advertisements.
Finally, these highly negative impacts on local broadcast station audiences and subsequent lower revenues caused by the importation of distant signals through new technologies are not merely speculative paranoia of broadcasters. This situation is eerily similar to the situation with respect to the satellite services illegally importing distant network signals into their local television markets in recent years. Like the nascent Internet streaming companies proposing this new service, these satellite services also started out small but quickly grew to major players in the video delivery industry. Until stopped by multiple lawsuits, their importation of these distant signals led to significant audience diversion and severe strains on the financial ability of many local television stations, particularly in a number of mid-sized and small markets, to serve their local communities. For example, in Missoula, Montana even by February 1998 the satellite delivered distant CBS affiliated station had an audience almost one-eighth (12.2%) the size of the local CBS affiliate, reducing the potential audience size and revenues of that local television station. Expert Report of Paul Bortz at 12-14, CBS Broadcasting Inc. v. PrimeTime 24 Joint Venture, Civ. No. 96-3650-NESBITT (S.D. Fla. May 28, 1998).

A. Impact on station profits and reactions
It is precisely the stations in such markets having less revenue, which will see an immediate impact on their profitability, as most of their costs cannot be lowered quickly. Television stations in mid-sized and smaller markets are already straining to keep their stations profitable and to continue serving their local communities. For example, if the average station in markets ranked 101-110 experienced a 10% decrease in revenues, they would see their small profits be wiped out and see a loss of over $200 thousand. See Table 11, pp. 22-23, NAB/BCFM 2000 Television Financial Report. The actual impact of a 10% decrease in revenues would be an 8.5% decrease in profits, as the typical 15% sales commission would not have to be paid. From over one-quarter of these stations losing money, there will probably be over one half that will be in that position. Ibid., p. 23.
The immediate loss in revenues and profits, and the prospects for even less revenues in the future, will lead those local television stations to have to adjust their spending and investments. Many, as shown before, are in very precarious positions already and may have to lower some costs to adapt to the reduction in advertising revenues. Unfortunately, there are just a few areas where a local television station can reduce costs. One area is in the programming costs, both in the amounts they will pay for nationally syndicated programming and the amounts they allocate to local programming. Local television stations may cut back these expenses resulting in less attractive syndicated programming in the future, and, more significantly, diminished news programming. With the added reduction in local news audiences caused by this third party streaming of local television station, some local television stations may elect to abandon their local news programming entirely. Local news production, while potentially profitable, does involve a large investment and operating cost commitment. If revenues do falter, this financial commitment may no longer be made.
Another possible victim from this downward spiral in local station’s revenues may be stations’ local community service and outreach efforts. Once again, there are many fixed costs faced by local television stations and when forced to find areas of cost cutting, local community services may be one such area of discretionary costs that will be reduced.
The impact on local station audiences, if these cost cutting decisions are made, will be profound. Local stations understand the impact on their future audiences if such decisions are made. With less attractive nationally syndicated programming and lesser quality news programming, local stations may see further reductions in audiences, starting a negative downward spiral. Future expenditures on programming will also have to be contracted weakening the station’s earning potential even further.
As the producer of their local news and informational programming, the local station will see the immediate impact as a program supplier. Other program suppliers, including the parent network, may also see an impact from this diversion of audiences to the Internet streamed signal. With local stations unwilling to pay for the best syndicated programming, the owners of these programs may also have to cut back on their spending for this programming. New syndicators, seeing the weakening syndication marketplace, will elect not to even enter with new products.
As for the parent networks, they too will suffer as their local affiliates weaken in their ability to spend on local programming. National networks count on the strong local programming of their affiliates to attract viewers to these stations with the expectation that they will either see promotional messages for the network programming and/or stay on that channel and watch the network programming. If local stations reduce the quality of their non-network programming, there will be less awareness of the network programming and less potential viewers staying on the local television station when the network programming is aired.
II. CONCLUSION
For at least the 21% of the U.S. households who do not subscribe to pay multichannel video services, free-over-the-air local television stations serve as the primary means by which they receive video news, information, and entertainment. The continued health and vitality of these local television stations is dependent, in large part, to their ability to serve their communities with distinctive and market exclusive programming. This market structure with exclusive market areas is present in many different industries, and has long been viewed as a pro-competitive, pro-consumer industry practice. Exclusive market areas for local television stations are one such example and provide the financial foundation for their local service.
Any change to this existing market structure may have tremendous effects on the ability of these local television stations to serve their local communities. Audience ratings and advertising revenues will be lower as audiences are diverted from some of the stations’ most popular programming. This reduction in audience ratings, advertising revenues and resulting profits may be especially significant for stations located in mid-sized and smaller markets, whose profitability is especially vulnerable. These local television stations have already recently seen their comparative advantage of providing mass audiences being eroded. Any further erosion caused by a third party streaming distant television stations into their markets may be especially harmful.
The impact may also be harmful to local television stations if viewers start to time-shift their viewing, disrupting the entire network program flow and negatively affecting the viewing to locally produced programming. Once again, with lower audiences local television stations may see lower advertising revenues, both immediately and over a longer term.
These lower advertising revenues of local television stations will of necessity lead to lower expenditures on programming, both what they spend on nationally syndicated and on local programming including local news. This cut back could start local television stations on downward spirals, with the result being a severe reduction in the quality of programming and service provided to their local communities.
The reduction in the amount spent on nationally syndicated programs and other actions related to the airing of network provided programs might have a subsequent negative effect on these program suppliers. They too may have to adjust and reduce their expenditures, thereby reducing the amount and quality of the programming they supply.
The end result of this cycle will be a weakening of the ability of a local television station’s ability to serve its local communities. With the present exclusive market areas for their programming, local television stations can prosper and invest in other programming that serve these communities. By tampering with that structure with the importation via a third party streaming local television stations, you threaten that successful provision of service.



Curriculum Vitae
Mark R. Fratrik

Vice President
BIA Financial Network
15120 Enterprise Court, Suite 100
Chantilly, VA 20151
703-818-2425
Mfratrik@bia.com
Education
Ph.D., 1981, Economics, Texas A&M University, College Station, TX
M.S., 1978, Economics, Texas A&M University, College Station, TX
B.A., 1976, Mathematics and Economics (honors), State University of New York at Binghamton
Professional experience
2001 – Present
BIA Financial Network
Vice President
III. Consulting in litigation and tax-related cases
IV. Developing of new broadcasting and related industry research offerings
V. Speaking at industry forums
1985 – 2000
National Association of Broadcasters
Vice President/Economist 1991 – 2000
VI. Supervised the Research and Planning Department.
VII. Conducted primary research about the broadcasting and related industries, used for testimony before the Congress and in filings at the FCC and other governmental agencies.
VIII. Conducted research and studies included in publications and reports distributed by NAB.
IX. Presented results of primary research and other analyses at industry forums.
Director of Financial and Economic Research 1985 -- 1991
X. Supervised the collection and dissemination of the annual industries financial reports
1980 – 1985
Federal Trade Commission
Bureau of Economics
Staff Economist
XI. Conducted analysis of proposed mergers and other arrangements.
XII. Conducted analyses of industry practices to evaluate economic impact.
XIII. Participated in litigation support in several antitrust cases.

Professional activities
American Economic Association – member
Southern Economic Association – member
Journal of Media Economics – reviewer
Journal of Broadcasting and Electronic Media - reviewer
Articles
“The Party’s Not Over, The Band is Just Taking a Break: How Radio Will Fare,” The Financial Manager, April-May 2001, pp. 29- 31.
“Broadcasting Industry Responses to New Technologies (with Rick Ducey), Journal of Media Economics, Fall 1989, Vol. 2, No. 2, pp. 67-86.
“Dual Distribution as a Vertical Control Device,” (with Malcolm B. Coate), Journal of Behavioral Economics, Spring 1989, Vol. 18, No. 1, pp. 1-19.
“The Myth of the Roaring 70s and the Quiet 80s,” NAB Info-Pak, November 1988.
“The Television Audience-Revenue Relationship Revisited,” Presented at the Broadcast Education Association Conference, April 28, 1986.
“How Important is Local Advertising to Today’s Television Station,” Broadcast Financial Journal, April 1986, pp. 12 –15.
“Predatory Pricing Theory Applied: The Case of Supermarkets vs. Warehouse Stores,” (with Richard Craswell), Case Western Reserve Law Review, Vol. 26, No. 1, 1985-86.
“Unanswered Questions About Franchising: Comments,” (with Ron Lafferty), Southern Economic Journal, 1984, pp. 928-932.
Books and Reports
State of The Radio Industry:
Radio Transactions 2000, 2001, BIA Financial Network.
Radio Industry Revenues – 2000 & Beyond, 2001BIA Financial Network
Ownership and Consolidation 2001, 2001BIA Financial Network
State of The Television Industry:
Television Transactions 2000, 2001, BIA Financial Network.
Television Industry Revenues – 2000 & Beyond, 2001BIA Financial Network
FM Subcarrier Market Report/Technology Guide (with David Layer), 1997, National Association of Broadcasters.
These Taxing Times: A Tax Guide for Broadcasters (editor), 1996, National Association of Broadcasters.
Strategic Planning Handbook for Broadcasters (with Richard Ducey), 1994, National Association of Broadcasters.
1994 FM Subcarrier Market Report, (with Kenneth Springer), 1994, National Association of Broadcasters.
The 1993 Tax Act: What it Means (editor), 1993, National Association of Broadcasters.
Fair Market Value of Radio Stations: A Buyer’s Guide, 2nd edition (with Bruce Bishop Cheen), National Association of Broadcasters, 1990.
RadiOutlook: Forces Shaping the Radio Industry (with John Abel & Richard Ducey), April 1988, National Association of Broadcasters.
Targeting Radio’s Future: Radio ’87, (with John Abel & Richard Ducey), September 1987, National Association of Broadcasters.
The Small Market Television Manager’s Guide (editor), 1987, National Association of Broadcasters.
Tax Reform: Effects on Broadcasters and Broadcasting (editor), 1987, National Association of Broadcasters.
“The New Audio Marketplace: Challenges and Opportunities for Broadcasters,” (with Richard Ducey) NAB Special Report, September 1985.
“The New Audio Marketplace: Challenges and Opportunities for Broadcasters,” NAB Special Report, September 1985.
Policy Research
“Television Web Site Activity,” Attachment 1, NAB Comments in re FCC examination of Disclosure Requirements for Television Broadcast Licensee Public Interest Obligations, December 2000.
“Independent Radio Voices in Radio Markets,” “Format Availability after Consolidation,” & “Interference from Low Power FM Stations to Existing Stations” (with David Wilson), Attachments A, B, and Volume 3, respectively, NAB Comments in re FCC examination of the Creation of a Low Power Radio Service, August 1999.
“Media Outlets by Market – Update,” “A Financial Analysis of the UHF Handicap,” Appendices A and C, respectively, NAB Comments in re FCC 1998 Biennial Review of Commission Ownership Rules, July 1998.
“The Television Industry’s Provision of Closed Captioning Services in 1996,” Attachment 1 NAB Comments in re FCC examination of Closed Captioning and Video Description of Video Programming, March 1996.
“Radio Station Financial Picture,” Attachment 13, NAB Comments in re FCC Establishment of Rules and Policies for the Digital Audio Radio Satellite Service, November 1995.
“The 1990 Children’s Television Act: A Second Look at Its Impact” (with Richard V. Ducey), Attachment 1, NAB Reply Comments in re FCC examination of Children’s Television Programming Rules, October 1995.
“The 1990 Children’s Television Act: Its Impact on the Amount of Educational and Informational Programming,” Attachment 1, NAB Comments in re FCC examination of Children’s Television Programming Rules, June 1994.
“Minimum Number of Owners under NAB Proposed Ownership Rules,” Appendix D, NAB Comments in re FCC examination of Revision of Radio Rules and Policies, May 1992.
“National Ownership Concentration of Television Stations,” Appendix A, NAB Comments in re FCC Review of the Policy Implications of the Changing Video Marketplace, November 21, 1991.
“AB Switch Availability and Use,” Attachment 1, NAB Comments in re FCC Examination of Carriage of Television Broadcast Signals by Cable Television Systems, September 23, 1991.
“FM Station Financial Picture,” Appendix B, NAB Request for Temporary Suspension of New Commercial FM Stations Allotment and Application Processing, February 10, 1991.
“Financial Analysis of Program Duplication for Radio Stations,” Appendix E, NAB Comments in re FCC Review of the Technical Assignment Criteria for the AM Broadcast Service, November 1990.
“Programming Aspects of the Territorial Exclusivity Rule,” “Financial Condition of Small Market Network Affiliated Television Stations,” Appendices A and E, respectively, NAB Comments in re FCC examination of Program Exclusivity Rules, January 1989.
“License Renewal/Transfer Study,” (with Michael Fitzmaurice), Appendix A in re FCC examination of Formulation of Policies & Rules Relating to Broadcast Renewal Applications, October 14, 1988.
“An Updated Examination of Market Concentration in Radio Markets,” Appendix E, NAB Comments in re FCC examination of Broadcast Multiple Ownership Rules, June 1987.
Testimony at the Environmental Protection Agency: In the Matter of Public Hearing on Federal Radiation on Protection Guidance: Proposed Alternatives for Controlling Public Exposure to Radio Frequency Protection, September 22, 1986.
“FM Facilities Reclassification Survey: Class B and Class C FM Stations,” (with Rick Ducey) Appendix A, NAB Comments in re FCC examination of FM Station Reclassification, August 1986.
“Financial Information on Commercial Radio Stations for AM Band Expansion Report,” Report V, submission of the Subgroup of Radio Spectrum Allocations on the Advisory Committee on Radio Broadcasting, May 1985.





APPENDIX B

Shortcomings and Challenges in the Restriction of Internet Retransmissions of Over-the-air Television Content to Canadian Internet Users
Benjamin Edelman


Introduction 2
Purpose and Summary 2
The theory of operation of Internet-based geographic access control systems 3
Problems in the production and use of geographic analysis tools 4
The lack of an established listing of Internet device locations 4
The difficulty of inferring device locations from Internet architecture 5
The special challenge of distinguishing between Canadian and American Internet devices 6
The lack of independent verification of geographic analysis system results 7
The offsetting effect of the need to avoid "false negatives" 8
Use of readily available methods to circumvent geographic analysis systems will be even more likely to thwart attempts to restrict distribution of high-value content 8
Proxy Servers 9
Other protocol-level mechanisms of circumventing geographic analysis tools 11
Complete bypass of the HTTP-based security system 12
Trends and Future Developments 13
Conclusion 14

XIV. Introduction
My name is Ben Edelman. My office address is (address removed – adresse enlevée). I have personal knowledge of the matters set forth in this Declaration.
I am a senior at Harvard College, and I work for the Berkman Center for Internet and Society at Harvard Law School as a systems administrator and multimedia specialist. I have previously worked as a computer purchasing consultant, a network designer and systems integrator, a custom software designer, a database specialist, and a designer of database-driven web sites. I have personal experience with RealMedia and Windows Media technologies, the primary methods used to transmit streaming video content over the Internet.
My experience includes six years as an Internet web server administrator. I have operated my own servers for six years, including a server receiving more than 10,000 hits per day. In addition, I have seven years of experience with the TCP/IP protocol on which the Internet is based, including six years administering TCP/IP-based networks. In my experience as a webmaster and network administrator, I have been asked to review log files for a number of purposes, including determining the geographic origins of individuals and groups of users.
In the course of my research at the Berkman Center for Internet & Society, and in the course of work performed for multiple clients, I have become familiar with automated systems that attempt to determine the location of an Internet user. In particular, I have reviewed the Digital Island Traceware technical documentation and have tested the Traceware system; I have reviewed the limited documentation available from bordercontrol.com and have tested the bordercontrol.com system; I have reviewed the Quova GeoPoint technical documentation. I am also generally familiar with DigitalEnvoy NetAcuity, InfoSplit OneToOne, and the RealMapping product line.
XV. Purpose and Summary
I have been asked by the National Association of Broadcasters to identify and describe limitations and challenges faced by Internet retransmitters of over-the-air television content in attempting to limit the transmission of such content to end users located within Canada.
I have conducted my analysis from the perspective that there will be numerous Internet users in the United States who desire to gain access to the retransmitted programming. Given the means available to Internet users today, there will be several ways in which they may gain access, including errors in geographic analysis systems and multiple means of bypassing such systems. Based on my knowledge of geography-based access systems, the purposes for which they have been used, and the means they employ to provide their services, it is my firm opinion that if over-the-air television programming is retransmitted over the Internet, users located in the United States will be able to gain access to the retransmitted programming.
XVI. The theory of operation of Internet-based geographic access control systems
To the best of my knowledge, commercial Internet-based geographic analysis tools have been available since no later than 1999, and there are today at least half a dozen geographic analysis tools available for purchase or license. The underlying theory of operation of such systems is that they attempt to determine an Internet user’s location from the user’s Internet Protocol address (“IP address”), a numerical identifier associated with a device connected to the Internet. Geographic analysis tools make this determination by referring to a previously prepared database that purports to identify the location of the device. Providers of geographic analysis tools use a variety of inferential methods to prepare these lists based on indirect information about location, including inspecting domain names associated with IP addresses, monitoring of routes by which data travels across the Internet, and determining the designated administrator of IP addresses assigned to relevant portions of the Internet. Digital Island Traceware Technical FAQ (<http://www.digitalisland.net/services/app_serv/faqs.html>), Quova Geopoint Technical Overview White Paper Historically, providers of geographic analysis tools have emphasized the use of their tools for demographic analysis, targeted content and advertising, native language content presentation, and product line segmentation. However, some providers of geographic analysis tools now suggest the use of such tools for the purpose of restricting content to users from particular geographic regions. Digital Island Traceware Technical FAQ (<http://www.digitalisland.net/services/app_serv/faqs.html>), Digital Island Traceware Brochure (<http://www.digitalisland.net/common/pdf/traceware_ds.pdf>), Quova GeoPoint API User Guide – Overview
Based on the materials I have reviewed from Digital Island, bordercontrol, and Quova, it is my understanding that an Internet retransmitter using a geographic analysis tool would design its access control system roughly as follows: Users visit the retransmitter’s web site seeking access to protected content, for which access is intended to be restricted to users within Canada. Upon receiving a request for access to such content, the retransmitter’s web server passes information about the user’s request (specifically, the user’s apparent IP address) to a geographic analysis engine embodied in software code provided by the selected geographic analysis provider. The geographic analysis engine attempts to find this IP address in its database, and if it does, it reports back to the web server the country it has associated with this IP address. Digital Island Traceware Technical FAQ (<http://www.digitalisland.net/services/app_serv/faqs.html>), Quova GeoPoint API User Guide – Overview, Quova Geopoint Technical Overview White Paper, RealMapping Technology Architecture If the reported country is in fact Canada, the retransmitter’s web server provides a web page giving access to the requested webcast content; if the reported country is other than Canada or is unknown, the retransmitter’s web server would refuse to provide such access.
XVII. Problems in the production and use of geographic analysis tools
The design of geographic analysis tools makes the process of offering access only to Canadian users highly dependent on the accuracy of the underlying database of IP addresses and corresponding geographic locations. However, the production and use of such a database face significant challenges that make the process prone to significant errors and failures.
A. The lack of an established listing of Internet device locations
Most seriously, there is no existing publicly-available centralized list documenting which specific IP addresses are associated with which specific physical locations. Regional Internet Registries (“RIRs”) Each Regional Internet Registry (or RIR) assigns IP addresses to entities located within its geographic area of the world. At this time, there are three recognized Regional Internet Registries, as documented on < http://www.aso.icann.org/rirs/>. The RIR responsible for Canada is also responsible for all of the Americas, Sub-Equatorial Africa and the Caribbean. maintain publicly-accessible records These publicly-accessible records are the so-called “IP-Whois system.” of the administrator of each IP range that has been assigned for use by computers, servers, and other devices connected to the Internet. However, individual assignments made by these high-level RIRs are often huge; a single assignment might allocate tens or hundreds of thousands of IP addresses that are ultimately assigned to devices in multiple countries. Nonetheless, for each range of addresses, the RIRs ordinarily publish only a single geographic location, typically associated only with the primary administrative contact for the IP address range, and not necessarily associated with all or even the majority of the devices using IP addresses within the range. Indeed, in the case of multi-city Internet Service Providers Multi-city Internet Service Providers include market leader America Online, whose internal network policies and use of proxy servers make its network especially misleading to geographic analysis tools seeking to determine the location of users’ machines. or multi-location companies, this information generally fails to identify the location of particular devices, since the RIR registration details address information for the ISP’s or company’s central headquarters, while the ISP’s customers or company’s users generally reside in many locations far beyond central headquarters. For example, if one were to seek to ascertain the location of an AOL user, publicly available information from RIRs would invariably point to Virginia, notwithstanding AOL’s customers across the United States and beyond. Thus, using information from an RIR’s publicly-available records about IP range registrations, it is not possible to draw a reliable inference about the location of a particular IP address. Digital Island Traceware Technical FAQ (<http://www.digitalisland.net/services/app_serv/faqs.html>)
Country-code top-level domains (ccTLDs) in the Internet’s Domain Name System (DNS) provide an alternative source of publicly-available information that might conceivably partially inform inferences about the geographic location of devices connected to the Internet. However, this data also entails numerous difficulties: While some ccTLDs enforce restrictions regarding who many register in their respective areas of the DNS, others allow open registrations by anyone. For example, the .TV and .MD ccTLDs openly encourage registrations by users located around the world. Even the Canadian .CA registry allows registration by non-Canadians: by associations with as many as 20% non-Canadian members, not to mention the Queen of England. http://www.cira.ca/official-doc/47.RPPG_00006EN.txt In addition, ccTLD registration policies change from time to time, are not always well-documented, and are not always consistently enforced.
Furthermore, even restrictions on registration of second-level domains within ccTLDs are insufficient to cause ccTLDs to provide reliable geographic information. For ccTLD registration restrictions ordinarily apply only to the identity of the registrant of the domain name; the underlying DNS and HTTP servers actually used to distribute content within that domain may be located anywhere on the Internet. Finally, domain names are ordinarily associated only with servers, not with the computers and other Internet-connected devices ordinarily operated by end users; thus, the domain name system does not inform analysis of locations of devices other than servers. In short, then, while ccTLD information can partially inform some assessments of geographic location, such inferences are not reliable and are not likely to inform analysis about the devices used by many end users.
B. The difficulty of inferring device locations from Internet architecture
Due to the shortcomings of publicly-available records regarding locations of IP assignments, geographic analysis providers also seek to use knowledge about the design and interconnections of Internet backbones to identify the location of a given Internet user. This method can sometimes partially inform attempts to draw inferences about the geographic location of particular IP addresses, Quova Geopoint Technical Overview, page 10 (“Analysis Process”) but such an approach is difficult and is likely to be unreliable in many instances.
One might try to infer the location of an IP address based on the segment of Internet backbone or interchange with which it appears to be associated. However, naming conventions of Internet backbone and interchange infrastructure, which could provide clues, are not widely standardized, reducing the ability of an automated system to draw inferences about the location of a device based on its name.
Alternatively, one might also try to determine location based on an analysis of the path by which data flows to and from a particular user. But “peering” and exchange relationships, which govern the passage of traffic from sector to sector and among Internet providers, are not always publicly disclosed and can fluctuate for reasons including technical efficiency, financial market conditions, human error, and system failure. These factors further reduce the reliability of a system that attempts to infer geographic location from the path that data seems to follow as it travels over the Internet.
Thus, it is difficult to draw reliable conclusions about the location of a user on the basis of traffic patterns to and from that user’s system.
C. The special challenge of distinguishing between Canadian and American Internet devices
Both the lack of a reliable locational database and the difficulties of inferring location based on Internet architecture are especially significant when attempting to differentiate Canadians from Americans on the basis of IP addresses. Many large Internet Service Providers offer access in multiple countries worldwide, and even small and medium-sized ISPs often offer service both in Canada and the United States. Similarly, the proximity of Canada and the United States and the economic ties between these countries cause many companies to operate offices in both. Moreover, both for ISPs offering services to the public and for multinational businesses operating their own networks, American and Canadian networks operated by a single entity are especially likely to be interconnected by dedicated private wide-area networks, precisely the design most troublesome for geographic analysis tools since such networks ordinarily have a single common point of connection to the public Internet. Quova Geopoint Technical Overview, page 11 (“Design Challenges – Corporate Network Proxies” For example, Nortel’s American operations almost certainly connect to the company’s Canadian headquarters via dedicated high-speed lines. Internet traffic that follows this same route within the corporate network might accordingly make users at the company’s American facilities appear to be located in Canada.
In addition, given the proximity of Canada to the United States, such transmissions are less likely to pass through well-known “peering points” (the connection points through which traffic flows between networks) than are other international communications, weakening the data-path means of analysis used by geographic analysis tools. Furthermore, while geographic analysis tools can look for telltale transoceanic delays in data transport between continents, there is ordinarily no such delay in communications between the United States and Canada, further reducing the ability of such tools to properly distinguish Canadian addresses from American ones. Thus, the peculiarities of the networks joining Canada with the United States are likely to make geographic analysis software particularly error-prone in this context.
Indeed, the increased causes of errors in differentiating between Canada and the US call into question certain statistics about the accuracy of geographic analysis software. In particular, claims that are based on a full sample of IP addresses worldwide are likely to overstate the accuracy of geographic analysis software when focused on the especially fine distinction between Canadians and Americans. If a company were to assert that its method is, for example, “98% accurate” on average across all its applications involving analysis of locations throughout the world, it is likely that the accuracy rate for Canadian and American location distinctions alone is lower than 98%, given the unique difficulties in this context, as described above. Without extended testing of actual geographic analysis tools, it is difficult to estimate the magnitude of this effect, but it seems possible that error rates could be as much as ten times greater in attempting to distinguish US users from Canadian users than in attempting to filter out Canadian users from the rest of the world.

D. The lack of independent verification of geographic analysis system results
Most commercial geographic analysis services do not publish their results, nor detailed information about their methods; this fact provides further reason to question the effectiveness of such tools. When an automated geographic analysis tool concludes that an IP address or range resides in a particular location, the company does not generally follow a practice of directly verifying its conclusion with the users of that IP address or range. Furthermore, leading providers of geographic analysis systems do not appear to publish their determinations of locations, nor make their determinations available on the web for free testing by interested Internet users. As a result, Internet users and system administrators have no opportunity to find or correct errors in their inferred locations. Indeed, without licensing access to each geographic analysis tool, an Internet user has no way to know whether each such tool determines his location accurately. Finally, for those users and systems administrators who wish to help facilitate improved accuracy of geographic analysis systems by submitting documentation about their network architecture, the web sites of leading geographic analysis tool providers reflect no formal way to do so, nor any request or suggestion that administrators might wish to do so.
E. The offsetting effect of the need to avoid "false negatives"
In the context of the intended application here – distinguishing Canadians from non-Canadians – the problem of accuracy becomes especially difficult due to the need to balance multiple and necessarily opposing types of errors, each of which presents serious problems. On one hand, the tools must avoid characterizing a user as Canadian when in fact she is not; errors in this regard are “leakage,” and it is only to address the leakage problem that geographic analysis tools are proposed at all. On the other, the tools must avoid characterizing a user as non-Canadian when in fact she is located in Canada; errors in this regard lead to user complaints, increased customer service expenses, and a negative impact on the image of the retransmitter as well as of the network providing the content being retransmit.
In the language of statistics, the problem here is the “type one versus type two error tradeoff”; in common language, geographic analysis tools are struggling to minimize both the incidence of false positives (erroneous permission of access) and false negatives (erroneous denial of access). Intuitively, being more certain that each person allowed in is actually in Canada means refusing access to more people who are possibly, but not certainly, in Canada.
Content delivery systems using geographic analysis tools as access control tools inevitably must make an explicit or implicit decision regarding the acceptable relative frequencies of each type of error. That is, a retransmitter company might resolve to reduce leakage by allowing retransmissions only to users who the geographic analysis tools found with the highest degree of confidence to be in Canada, but in doing so, the number of erroneous denials of service would unavoidably increase. With a given level of technology, it simply is not possible to reduce both types of error simultaneously. Furthermore, the commercial incentives of an advertising-driven business model strongly disfavor false negatives, causing still greater impediments to attempts to minimize false positives.
XVIII. Readily available methods to circumvent geographic analysis systems will likely thwart attempts to restrict distribution of high-value content
The technical and practical issues discussed above would make it difficult to operate a reliable geographic location access system even if there were no substantial incentives for gaining unauthorized access. But in the present context, the powerful incentives for gaining access add even more substantial impediments to a successful geographic location analysis system.
After reviewing the web sites of leading producers of geographic analysis tools, it is my understanding that these companies have focused much of their promotional efforts on applications in which users have little or no incentive to circumvent or attempt to circumvent the geographic analysis system.
For example, when a geographic analysis system is used to distribute targeted advertising intended only for viewers in a particular region (say, where a particular product is available), Digital Island Traceware Brochure (<http://www.digitalisland.net/common/pdf/traceware_ds.pdf>), The Advantages of RealMapping typical content providers are not motivated to completely prevent “spillover” beyond the targeted audience. Neither are end users typically motivated to attempt to receive the advertising intended for those in other locations.
There is similarly little incentive for end users to circumvent geographic analysis systems when such systems are used to transmit web pages in a user’s likely primary language. Digital Island Traceware Brochure (<http://www.digitalisland.net/common/pdf/traceware_ds.pdf>) When such systems are used for demographic analysis, Digital Island Traceware Brochure (<http://www.digitalisland.net/common/pdf/traceware_ds.pdf>) there is ordinarily neither such an incentive nor an obvious opportunity to do so.
However, the present context is notably different, in that access to retransmitted over-the-air television content is likely to be valuable to certain users, who may therefore be motivated to take affirmative steps (or even to expend considerable effort) in attempting to receive such transmissions. Accordingly, it is important to consider the means by which they might do so as well as the likely extent of their success.
A. Proxy Servers
Since the initial development of the HTTP protocol used by the world-wide web, certain users and networks have come to rely on “proxy servers,” Internet devices that receive and forward certain types of content requests between a user's own server and a web site she wishes to view. Such use has multiple motivations: A user who values her privacy might find that a proxy server can prevent web server operators from learning a variety of facts about her identity. A network operator concerned about bandwidth expense or inappropriate use of Internet connectivity might use a proxy server to improve network efficiency or to restrict access to certain sites.
However, via the use of proxy servers, interested Internet users are able to disguise their location to most systems, including commonly-used geographic analysis tools. For example, an American Internet user might ask a Canadian proxy server for the web page <http://www.retransmitter.ca/channelname>. On behalf of the user, the proxy server requests the specified page from the specified server; then, the proxy server forwards the results to the user who submitted the initial request. In this case, the retransmitter’s geographic analysis tool would ordinarily be unaware that the destination of the requested page was not the proxy server itself but rather an unspecified end user; even if the geographic analysis tool recognized the proxy server as such, it would be unable to determine the actual location of the end user.
While public proxy servers are not widespread at this time, there is significant evidence that their deployment is increasing. Companies such as anonymizer.com, ZeroKnowledge, and SafeWeb have long provided free public proxy service, supporting this service with sales of related services and/or with interstitial or header advertising. More recently, peer-to-peer systems have begun to take advantage of the availability of widespread high-speed unmetered connections to offer widely distributed proxy server networks. SafeWeb Triangle Boy (<https://fugu.safeweb.com/sjws/solutions/triangle_boy.html>) For example, in some implementations, public-spirited users with unlimited high-speed Internet access volunteer the use of their system, perhaps when not otherwise in use, to relay content requested by others, most likely by Internet users who for any of several reasons cannot retrieve that content directly from its source. In these peer-to-peer proxy server systems, requests forwarded by a volunteer proxy server seem to be from the volunteer herself, despite that the resulting page will ultimately be seen at a different address, in general at a different geographic location. There is no readily-available practical method by which web servers or geographic analysis systems could prevent accesses via a distributed peer-to-peer network of this type; ordinary blocking methods (for example, denying all accesses of apparent origin at anonymizer.com) would be ineffective in the face of thousands proxy servers at dispersed and constantly-changing addresses. Indeed, precisely because this proxy implementation is so effective, recent news articles reflect that the United States government may commission the availability of such a system for use by interested Chinese Internet users who seek to circumvent the Chinese government’s Internet filtering systems. “U.S. May Help Chinese Evade Net Censorship,” New York Times, August 30, 2001 (<http://cyberatlas.internet.com/big_picture/geographics/article/0,1323,5911_51151,00.html>)
The availability of exclusive high-value content protected by geographic analysis systems would be likely to encourage additional efforts at circumvention via proxy servers. At the moment, the primary benefit to American Internet users American Internet users continue to represent the bulk of the Internet population. “The World’s Online Populations,” Internet.com Study (<http://cyberatlas.internet.com/big_picture/geographics/article/0,1323,5911_51151,00.html>). The same applies to Internet users from all other countries, save for those that filter the Internet (China, for example) and those that contemplate permitting access to retransmit network television content (Canada, for example). of Safeweb’s peer-to-peer system is anonymity in web browsing, but if Internet retransmissions of over-the-air television content were accessible to Americans only if they used Safeweb (or a similar system), use of such systems would likely rise dramatically. In such circumstances, it is also likely that Canadian Internet users, and those non-Canadians with IP addresses wrongly classified as Canadian by leading geographic analysis tools, would offer their services as proxy servers to friends, acquaintances, and perhaps even strangers. For example, Canadians operating proxy servers might auction the use of their servers via ebay or a similar system, or they might sell lists of access mechanisms as is currently the case in the realm of Internet pornography. In short, then, were over-the-air television content available via Internet retransmissions and restricted by geographic analysis tools, it is likely that proxy servers would see increasing use in effectively bypassing such restrictions.
B. Other protocol-level mechanisms of circumventing geographic analysis tools
In addition to proxy servers, there exist multiple alternative methods of purposefully circumventing the protections of geographic analysis tools.
For example, tunneling methods Among tunneling methods are PPTP, IPSec, and all other means of providing “virtual private networks” (VPNs). involve repackaging entire IP packets so as to send them to their destination via a remote tunneling server, thereby hiding a user’s actual location and causing the user to appear to hold the IP address of her tunneling server. These methods are widely deployed in the context of corporate networks, and the end user’s necessary client software is included with recent versions of the Microsoft Windows operating system.
Similarly, terminal services methods Among terminal services methods are Microsoft Windows Terminal Server, Citrix MetaFrame, X Windows, and Symantec pcAnywhere. involve transmission of keystrokes, mouse movements, and display information between end user and a remote terminal server. Here again, geographic analysis systems would base their findings on the IP address of the terminal server, while the end user is located elsewhere.
The net effect of all these methods is that an interested and determined user has access to multiple methods by which to appear to be located in Canada, for the purpose of receiving over-the-air television content from an Internet retransmitter. Furthermore, such methods are proliferating as proxy servers and tunneling methods expand due to the rise of privacy concerns and the proliferation of corporate remote-access systems.
C. Complete bypass of the HTTP-based security system
Discussion so far has considered various means by which interested Internet users might provide a geographic analysis tool with an IP address unlikely to be associated with the actual location of the end user. However, there exists an altogether separate method of circumventing geographic analysis systems used for the purpose of restricting access to streaming media content: Interested users might link directly to the streaming media content, without first passing through those web servers secured by geographic analysis tools.
Indeed, current geographic analysis tools are generally intended to be integrated into a content provider’s system at the level of the HTTP server (“web server”), and they are ordinarily used precisely and only in this way when used to customize a web site. Digital Island Traceware Technical FAQ (<http://www.digitalisland.net/services/app_serv/faqs.html>), Quova GeoPoint API User Guide, RealMapping Technology However, an HTTP server is not used to deliver streaming media under the market-leading implementations of RealVideo and Windows Media Architecture streaming video; rather, these systems use proprietary streaming media server systems, namely the RealServer and Windows Media Server, respectively. To the best of my knowledge, there is no off-the-shelf method of integrating geographic analysis tools with these streaming media servers, and based on my review of the technical documentation of these two servers, as well as informal conversations I have had with commercial webcasters who have attempted similar tasks, I have concluded that it would be extremely difficult to perform such integration even with custom software code.
As a result, in most implementations, it is likely to be possible to “deep link” directly to the desired streaming video content without first accessing the HTTP server that ordinarily provides the link to the streaming media content; in this way, it is possible to bypass the sole point at which geographic analysis access controls are applied. It may be difficult for novice users to ascertain the location required for “deep link” access directly to streaming video content. However, this process need only be performed once per channel on each retransmitter. Thereafter, straightforward instructions could be written by a single user and posted for use by all. Furthermore, once the destination of the deep link to a given over-the-air television channel becomes well-known, it is straightforward to create a link to that channel from any web server, and this link alone would be sufficient to grant access, without any additional instructions or access procedures. In January 2000, Streambox.com created precisely such a link to over-the-air television content retransmitted by iCraveTV.
For a retransmitter unable to attach geographic analysis software to its streaming media server due to the lack of appropriate interfaces on the streaming media server, such deep linking is difficult or impossible to prevent. Incoming requests that result from deep links appear the same to a streaming media server as do incoming requests that result from clicking through a geographic analysis tool’s authentication system on the retransmitter’s web site. Thus, in such circumstances, it is likely to be impossible to prevent deep linking to streaming media content without simultaneously preventing authorized access.
XIX. Trends and Future Developments
Looking forward, there are significant reasons to expect it to become harder, not easier, to produce accurate geographic analysis tools and to thereby restrict retransmitted over-the-air television content to a Canadian audience.
As noted above, the accuracy of geographic analysis systems – which is substantially impeded in the first place by the lack of reliable information about the location of the devices identified with particular IP addresses – is further hindered by the rise in deployment of proxy servers, tunneling systems, and terminal services. Such systems can cause geographic analysis systems to draw erroneous conclusions about the locations of end users; thus, their increased use reduces the accuracy of geographic analysis tools.
Furthermore, the increased deployment of mobile network devices may make it more difficult to determine the location of an Internet user at the time of access to a particular site, likely differing from the fixed location of the user’s service provider.
Finally, geographic analysis tools are likely to suffer in effectiveness due to the increasing availability of automated tools and generally-known methods for bypassing security systems. For example, the Bitbop Tuner http://www.bitbop.com allows users to listen to a streaming radio station over the Web without ever visiting that station’s web site – a method that would most likely circumvent security implemented at the level of the web server. Other recent tools suggest additional methods for novice users to deep link around HTTP-level security systems, while multiple Internet discussion boards periodically consider this topic from time to time. Thus, there is substantial reason to believe that such methods and skills will become increasingly widespread and well-known in the future.
XX. Conclusion
Given the current development of geographic analysis tools and given the systems and techniques available to Internet users today, there will inevitably be multiple ways in which Internet users may gain access to any over-the-air television content retransmitted over the Internet. The interconnected global architecture of the Internet is at odds with an attempt to “fence off” a country through technological means, and the challenges confronting geographical access control systems that attempt to do so are, in my opinion, effectively insurmountable at the present time.