The Internet: Access Avenue for Videoconferencing

Ramiro Calvo <ramiro@apple.com>
Apple Computer, Inc.
1 Infinite Loop, MS 301-4H
Cupertino, California 95014, USA

Venilde Jeronimo <jeronimv@cgs.edu>
Center for Politics and Economics
Claremont Graduate School
170 East Tenth Street
Claremont, California 91711, USA

Abstract

The affordability of multimedia personal computers, the expanding use of local- and wide-area networks, and the emergence of digital audio and video technologies that make possible transmission of voice and images over existing computer networks are key developments that now allow videoconferencing (VC) systems to become mainstream. Of key importance for the widespread use of VC is the increasing acceptance of the Internet as a new paradigm for communications. This leads us to suggest that the growing global connectivity available through the Internet provides the necessary access avenue for voice and video communications to become mainstream. Furthermore, we suggest that VC will be a sought-after access tool and will make possible a new generation of Internet features for a wide range of uses and users in the commercial and noncommercial sectors. The recent proliferation of Internet voice and video providers and Internet congestion are also addressed in this paper, as well as the question of whether universal service should be redefined to include access to the Internet.

1. Introduction

A century and a half has passed since 1851, when Nathaniel Hawthorne envisioned the information age as upon us [1]:

It's a fact ... that, by means of electricity, the world of matter has become a great nerve, vibrating thousands of miles in a breathless point of time? Rather, the round globe is a vast head, a brain, instinct with intelligence or, shall we say, it itself a thought, nothing but thought, and no longer the substance which we deemed it!

By the early 1960s, Marshall McLuhan used the term "global village" to denote worldwide communication through electronic systems, which he prophesied would lead to a coalescence of human awareness into a single global community [2]. Today, an information society is said to be upon us in which the saturation and importance to daily living of telephones, media systems, and computer networks are characteristic of our increasingly technological lifestyles [3].

The growing need to support technical and social activity that occurs across geographic distances in this information age, however, has not been fully satisfied by the current technologies of phones, faxes, electronic mail, and videoconference rooms [4]. Videoconferencing (VC) systems have been promoted as a technology to support remote collaboration since AT&T introduced the Picturephone in 1964 (Table 1), yet this vision has not been fully realized. A study conducted by Johansen, Hansell, and Green, addressing the question why teleconferencing was not implemented in business environments on the scale envisioned in the early 1970s, suggests that the optimistic demand for forecasts was based on the assumption that readiness factors (e.g., economic and social trends, technological developments) would rapidly evolve into enabling forces (e.g., user-oriented systems, support services, effective user networks) [5]. The Johansen et al. study focused on the use of videoconference rooms in business, but recent technical and infrastructural developments prompt the adoption and use of VC systems on a much larger scale than previously imagined.

Table 1. History of videoconferencing (VC) technology

Source: Modified from Wall Street Journal (27 February 1996).

First, technological advances have made the cost of VC plummet. In the 1970s, color cameras the size of a two-drawer file cabinet required studio lighting and sophisticated daily setup and maintenance at a cost of $50,000 to $100,000. In the 1980s, conference environments used shoebox-sized cameras at a cost of $5,000 to $10,000. Today, for $200, an end user can acquire a desktop VC kit that includes a digital camera and software to be used at home or work.

Second, several developments have allowed VC applications to become mainstream: the affordability of multimedia personal computers, the expanding use of local-area networks (LANs) and wide-area networks (WANs), and the emergence of digital audio and video technologies that allow voice and images to be transmitted over existing computer networks. Of key importance for the expanding use of VC applications is the Internet--the network of networks--which is increasingly accepted as a new paradigm for communications [6]. We suggest that the growing global connectivity available through the Internet provides the necessary gateway for the widespread use of audio and video communications. By the end of 1995, there were an estimated 4,852,000 Internet hosts worldwide, with Internet access points growing at an estimated 50 percent per month and subscriber growth close to 30 percent per month.

The popularity of the Internet as a medium of communication has recently led the America's Carriers Telecommunications Association (ACTA), a trade association of independent long distance carriers, to file a petition with the Federal Communications Commission (FCC) to stop companies from selling software and hardware products that enable use of the Internet to supply voice long-distance and international services unless these companies are regulated as long-distance carriers. This policy development and its implications for Internet video communications, and the question of whether universal service should be redefined to include access to the Internet are briefly discussed in the last section of this paper. The following section defines VC as it is discussed here, as well as the key technological developments responsible for the recent VC revolution. There is a brief section on Internet voice and video communication providers versus telecom carriers and one on concerns about Internet congestion.

2. Videoconferencing technology

2.1. The infrastructure evolution

Videoconferencing has an array of definitions; here we use a simple one: VC is interactive and collaborative audio and video communication between two or more people in different locations. In some cases, collaboration may extend to sharing of documents and applications. This "revolution" is a result of the evolution of a variety of technologies that have been put in place for reasons other than VC. The development of multimedia personal computers and LANs and WANs has played an important role in the development of the VC market. We must first look into these technologies in order to better understand the potential of VC today.

The "dashboard" of the information universe, the personal computer, is playing a key role in the VC revolution. In recent years, multimedia PCs have become commonplace at work and in the home. Multimedia PCs are characterized by their ability to capture and play back audio and video data, their use of 100 MHz+ processors, and their built-in networking capabilities.

Apart from the new functions added to personal computers, the most dramatic change is the increase in processor power. In the early 1980s, PCs used a 6 MHz processor; today's computers are powered by 200 MHz processors with wide 32-bit buses, advanced floating-point units, and extensive use of pipelining. This increase in computational power has made possible software-only audio and video compressors. In effect, as the processing speed increases, so does the compression ratio, which in turn reduces network bandwidth requirements.

At the same time, built-in network interfaces have become commonplace. Many PCs come with 10-megabit Ethernet coupled with high-performance direct memory access (DMA) processors able to handle steady streams of digitized data. Computer peripherals such as monitors have also evolved. Not long ago, 12-inch text-only black-and-white (or green) monitors were state of the art. We now see 17- and 21-inch color graphics monitors capable of updating 1024 x 768 pixels at 75 times per second, offering much higher picture quality than a television set.

Another key requirement of VC is the availability of interconnected personal computers. When people first started using PCs, the required connectivity was a simple serial interface between the computer and its printer. Then small groups of computer users emerged, creating the need to exchange files among themselves through small LANs. When universities and mid-size to large companies required small workgroups to interact with each other, clusters of Ethernet LANs were tied together into larger campus networks with Fiber Distributed Data Interface (FDDI) backbone networks. Today's campus networks connect to each other, creating a network of networks, through dedicated or shared T1 and T3 leased lines.

With the success of the Internet, Internet service providers (ISPs) are now bringing this network of networks to the home. The result is that we now have a critical mass of computers worldwide connected to one another, just like the telephone system.

Not only is the entire world becoming interconnected, but the networks are getting faster. For example, not too long ago, connectivity between PCs consisted of point-to-point 300-baud modem connections. We now have worldwide connections via the Internet at 28,800 baud (V.34), 128,000 bps (ISDN), 1,544,000 bps (T1), and 45,000,000 bps (T3). With cable modems and asynchronous transfer mode (ATM) networks, megabit networks to the home could soon become yet another low-cost alternative for Internet users. At work, LANs have gone from 230 kbps with LocalTalk, to 10 Mbps with Ethernet, to 100 Mbps with Fast Ethernet, FDDI, and ATM. This increase in capacity has resulted from improvements in various technical areas: modulation techniques used in modems, encoding schemes for sending high-speed data across copper cable, the development of single-mode fiber optic cable used in cross-continental connections, and low-cost chip technology able to operate at gigahertz speeds.

With the availability of powerful personal computers and easily accessible digital networks, the basic infrastructure for inexpensive VC is now in place. The point of intersection between the required network bandwidth for the transmission of digitized audio and video data and the availability of low-cost network bandwidth represents a major paradigm shift. The paradigm shift for audioconferencing over the Internet occurred within the past year, and we believe a similar shift is now occurring for videoconferencing.

Unfortunately, the dramatic increases in available bandwidth and improvements in data compression will not be enough to prevent congestion on the Internet due to the explosion in the number of Internet users. We predict that aggregate bandwidth needs will most likely overtake technological progress. As a result, we believe ISPs will compete for customers by offering "uncongested" Internet access, creating a huge demand for gigabit networks. This can only be good news for telecom carriers that lease digital networks.

2.2. Inside videoconferencing software

VC solutions build on three fundamental building blocks: audio and video digitizers, fast microprocessors, and a built-in network interface. As mentioned before, all three building blocks are now considered basic functions offered in hundreds of thousands of multimedia PCs. With these building blocks, companies have been able to develop solutions that support point-to-point, multiparty, and broadcast conferencing systems. Since network protocol stacks and compressors/decompressors (CODECs) are implemented in software, VC systems can be easily upgraded over the Internet with algorithmic improvements. It is therefore critical to first build software infrastructures capable of handling a wide variety of network protocols and compression algorithms.

VC software is a union of two of the most challenging computer science topics: real-time processing and distributed processing. Unfortunately, the Internet and the PC were designed for store-and-forward transactions. Developers have therefore relied on raw computational power and simple buffering mechanisms to work around the lack of real-time network and operating system (OS) services.

Simple buffering mechanisms are used to smooth out the variable OS and network response time. To reduce increasing end-to-end delays caused by buffering mechanisms, many conferencing systems use protocols that move data in a best-effort manner with simple error detection algorithms that avoid end-to-end error recovery mechanisms.

Finally, algorithms for monitoring central processing units (CPUs) and network loads are used to dynamically determine the rate at which video data are transmitted. It is critical to avoid overloading CPUs and networks to ensure that existing e-mail and file transfer services are not affected. In addition, overloaded systems typically waste network bandwidth by transmitting data that may never reach or be used by the receiving side.

As the VC market heats up, development for real-time network and OS solutions will increase. With real-time services incorporated into next-generation networks and OSs, applications will be able to count on processor time for network protocols and software CODECs, which would reduce buffering requirements and end-to-end delays. With next-generation routers, efficient use of IP multicast will improve network bandwidth use by avoiding needless duplication of data in multipoint connections. Compression algorithms will continue to improve as the availability of CPU cycles increases. In summary, most solutions being developed are scalable, so they can take advantage of increased processing power, new protocols, and new network infrastructures.

2.3. Standards

A key factor in the success of the telephone and the fax machine has been the ability to communicate without regard to the type of equipment used at either end. The goal of the VC industry must be to have a "video dial tone" if it is to become truly ubiquitous. The most popular VC standard to date is H.320. which ensures interoperability between large boardrooms, roll-abouts, and PC-based VC systems. However, H.320 requires dedicated circuit-switched network services, such as ISDN or T1 lines. With the emergence of the Internet, which uses both circuit- and packet-switched networks, we will see the development of new standards more appropriate for this mixed environment.

Today's Internet VC products are based primarily on proprietary protocols, so they are noninteroperable. The industry is working toward a new set of standards that ensure interoperability in multimedia data formats, as well as for document sharing. The standards come from the Internet Engineering Task Force (IETF) as well as the International Telecommunications Union (ITU), as shown in Table 2.

Table 2. Related Internet videoconferencing standards

                       Industry
Function               Standard         Notes
--------------------------------------------------------------------------------------------
Audio input            Line-in signal   * 8, 11, 22, 44 kHz sample rate
                       level              8 or 16 bits per sample with mono or stereo inputs

Video input            NTSC             * National Television System Committee (used in 
                                          Americas and Asia)
                       PAL              * Phase Alternation Line (used in Europe, except in 
                                          France, where SECAM is used)

Audio compression      G.711            * 64 kbps, 2:1 compression with a 128 kbps source
                       G.722            * 64 kpbs, 4:1 compression with a 256 kbps source
                       G.728            * 16 kbps, 8:1 compression with a 128 kbps source
                       G.723            * 6.4 kbps, 20:1 compression with a 128 kbps source
                                        * 5.3 kbps, 24:1 compression with a 128 kbps source
                       GSM              * 13.3 kbps, 10:1 compression with a 128 kbps source

Video compression      H.261            * Tailored for 50 kbps to 1.2 Mbps data channels
                                          QSIF = 176 pixels x 144 pixels x 24 bits/pixel
                                          SIF = 352 pixels x 288 pixels x 24 bits/pixel
                       H.263            * Tailored for 15 kbps to 128 kbps data channels
                                          QSIF = 176 pixels x 144 pixels x 24 bits/pixel

Shared network         RTP              * IETF standard
protocols              H.323            * ITU standard

Collaboration          T.120            * Support for multipoint connections with shared
                                          whiteboard and file transfer capabilities

Application/document   None             * To date, only proprietary solutions offer 
sharing                                   this function
--------------------------------------------------------------------------------------------

Source: Authors' compilation.

2.4. What's next?

Despite the enormous progress thus far, much work remains. First, we challenge the Internet Society to provide a "universal" directory service for e-mail as well as IP addresses for VC for those who want to be published. Exchanging e-mail as well as VC IP addresses at social gatherings is becoming the pickup line of the '90s. This information is typically exchanged verbally, but what is really needed are electronic directory services. Some already exist, but they are not broad enough to be truly useful.

Second, we hope ISPs will provide static IP addresses to all subscribers. Currently, dynamic IP addresses make it difficult to create directory services (or phone books) for VC users. Can you imagine how useful (or useless) e-mail would be if your address changed every time you connected to the Internet? This is the case with dynamic IP addresses and VC. As the market matures, software that uses similar network protocols and consistent human interface designs will win customers over. For example, the human interface of the telephone has universally recognizable dial tones, ringing signals, and busy signals. We hope that Internet VC will be just as easy to use and just as universally compatible. Our prediction is that continuous Internet connections will soon become affordable. This will allow users to receive unscheduled VC calls much as they receive telephone calls today. Currently, home users must first connect to the Internet before being able to receive phone calls through the Internet. Until this changes, Internet conferencing will remain a hobby, instead of the true communications tool it promises to be.

Third, the biggest challenge is one of network bandwidth. The number of Internet users is increasing, the bandwidth requirements put on the network by each user are increasing, and the length of time users stay on the network is also increasing--all of which adds up to one very big global traffic jam. Fortunately, digital fiber optic networks have tremendous capacity. For example, a single cable containing 24 fiber optic strands can support approximately 5 million simultaneous videoconferences: n = sb/u, where n is the number of users (5 million), s is the number of strands per cable (24), b is the bandwidth per strand (6 Gbps), and u is the bandwidth per user (28.8 kbps).

Internet traffic, with VC as part of it, is finally generating the digital traffic volume for which most phone companies have been waiting. Bandwidth availability may become the main differentiator among private online services and ISPs. Today, these services are evaluated by "on-ramp access speeds" (e.g., 14.4, 28.8, or 128 kbps). In the future, we expect that they will also be evaluated by the backbone network bandwidth available.

3. Internet voice and video communications

Over the past decade, many people have experimented with the idea of bringing the telephone, with its real-time communications technology, to the Internet via the desktop computer. Several programs were released in the beginning of 1995 to support this concept; Apple Computer's QuickTime is an example. These programs do not represent a "half-and-half integration" with conventional telephony, but rather an entirely parallel and independent system [7]. Basically, they allow the Internet to be used to make computer-to-computer long-distance and international audio and video calls. These systems are being targeted for home use (e.g., for personal communications and online shopping) and are expected to become as standard as telephones and faxes are today [8]. Outside of home use, many companies are installing desktop VC systems in branch offices to allow executives and employees to connect to headquarters, other employees, or clients and customers. Phone and videophone connections can now be built into Web pages, which can be useful to the corporate world. For instance, users can have simultaneous face-to-face meetings and, if necessary, jointly view and edit a document on the Web. Educational facilities (e.g., for distance learning), nonprofit organizations, and government entities are also potential users of audio and video systems.

The advent of providers of Internet voice and video systems raises several questions, two of which we briefly discuss here. With the increasing popularity of the Internet as a communications medium, (1) is use of the Internet for voice and video communications a competitive threat to phone companies, and (2) how can ISPs deal with Internet congestion?

3.1. Internet phone and video providers versus phone companies: A simple cost calculation

Internet calling, which began as an amusing pastime a year ago, is now starting to be used by an array of users as a less expensive medium to transmit voice and video and bypass high long-distance and international charges. As the supply (production) of Internet audio and video communications systems increases, the cost of production per unit falls (assuming some learning curve exists). A cost difference appears to exist between telephone companies and Internet phone providers; it is likely to decrease further over time. The cost to a user of Internet telephony is basically the cost for the Internet connection. Unlike phone companies, which charge according to time and distance (the longer you talk and the farther you call, the more you pay), Internet connections (even though they travel the same wires) are sold at flat monthly rates, and the connection to an ISP is usually a local call. For instance, ISPs may charge users approximately $10 for five hours of access and then approximately $3 for each additional hour. Five hours is 300 minutes, divided by $10 is 3.3 cents per minute. The average residential long-distance call costs about 22 cents per minute--seven times as much. Even with additional hours at $3 each, the ISP costs only 5 cents per minute, still much less than the 22 cents charged by telephone companies.

Several problems can be identified with Internet calling. First, lack of interoperability imposes a cost. The cost implied by interoperability disappears within just a few years, however, due to faster growth of the industry with interoperable components. Although costs begin at a higher point, production moves quickly down the learning curve, resulting in accelerated growth and lower prices compared to proprietary systems [9]. Interoperability is a result of the economic incentives for cooperation that exist among suppliers who are induced by the positive network externalities associated with being on the Internet. Suppliers of Internet voice and video services are multiplying, with many collaborating on interoperable protocols, and third-party developers are also creating the necessary protocols to connect different systems.

Second, a user must have a sound card and a modem, although these are now standard on most personal computers, which are increasingly affordable.

A third drawback to Internet calling is poor quality and convenience (all calling parties must be logged on); however, no one disputes that technology is improving in this area.

Nevertheless, current demand seems to be optimistic. An estimated 1.5 million Internet phone products have been shipped [10]; 20,000 individuals now use Internet phone software regularly [11]; and 2,000 downloads are being reported by WebPhone alone [12]. By the end of the year, it will not be a surprise to have Internet telephony as a standard feature in Web browser programs and videophone connections built into many Web pages. Call-back systems are another area in which the Internet can be used to bypass and compete with phone networks.

3.2. Internet traffic jams

As mentioned earlier, the biggest challenge is network bandwidth. While digital networks have tremendous capacity, there is concern about Internet congestion, especially with the increased demand for voice and video communications [13]. With Internet traffic jams, users have begun to experience delays, and some of these users may pay more to circumvent jams. This has led the economic community to characterize the pricing structure of the Internet as suboptimal. Three options to price the Internet for congestion control and interconnection are being discussed: (1) flat-rate pricing, (2) use-sensitive pricing, and (3) transaction-based pricing [14]. The first option refers to the fee users pay today to connect, regardless of the amount of information sent; the second and third options are new.

With use-sensitive pricing, users pay a portion of their Internet bill for a connection and a portion for each bit sent or received. Options could include use-sensitive pricing during peak hours and flat-rate pricing off peak. As for the transaction-based pricing option, it is determined by the characteristics of the transaction and not by the number of bits. Some economists argue that, unlike the telephone system, it is difficult for the Internet to support transaction-based pricing because of its distributed computing environment; on the other hand, interconnection points with flat-fee settlements may provide improper incentives for use and may lead to inequitable cost recovery by the interconnected firms [15]. Use-sensitive pricing has recently been introduced by Internet access providers such as BBN, MCI, and Uunet [16]. There does not seem to be a consensus on which pricing option is best for congestion and interconnection problems. There does appear to be a consensus, however, that not all traffic on the Internet needs to be treated the same, and therefore service qualities can affect resource allocation and pricing [17]. This view supports pricing based on service quality as possibly the only way to limit the demand on the guaranteed service quality.

The National Science Foundation (NSF) appears to favor some sort of prioritization of traffic over the Internet [18]. Recently, NSF introduced a new angle to its connections program associated with its concern over congestion. NSF will provide grants to college and university campus network service providers for applications that require high-performance networking. The hope is that the grants will stimulate the development of a technological option for the Internet to provide a new way of connection to give a guaranteed level of service at a national level. NSF's connections program is expected to spur the development of switches and routers to alleviate information bottlenecks. One solution might include prioritization of traffic on the Internet. This is analogous to postal services. To send a package, an individual has several options: overnight mail, first-class mail, or third-class mail. The cost of the package delivery depends on its designation. Another solution might involve diverting specially coded traffic to high-performance special-use networks, such as NSF's vBNS (very high speed Backbone Network Service).

4. Public policy issues

For several decades, telecommunications worldwide were regulated through a public utility paradigm, under the economic auspices of a natural monopoly and a public interest rationale (universal service) [19]. Today, for technical and economic reasons, telecommunications policy is undergoing transition and reevaluation in most countries. The telecommunications sector is being exposed to deregulation, liberalization, and privatization. Many countries are experimenting with flexible decentralization, which suggests that new competitors (many based on computer, mobile, or broadcast technologies) should be allowed to compete to provide a broad range of services, from telecommunications infrastructure to information services.

One medium that is only now receiving attention in policy circles is the use of the Internet for audio and video communications. While the global trend is toward deregulation, a host of issues are now surfacing in regard to whether the Internet needs government oversight:

All these issues merit attention. Here, we deal with the first and last issues because they are recent issues of public debate.

4.1. Regulating Internet voice and video communications

Currently, long-distance companies are regulated by the FCC and state regulatory commissions. They are required to pay access fees, file tariffs in order to provide long-distance telephone services, and contribute to the Universal Service Fund to extend phone service to the general population. On 4 March 1996, ACTA (a trade association founded in 1985 by independent long-distance companies to serve the needs of small businesses) filed a petition [20] aimed at companies and software developers that provide voice services via the Internet but are not regulated by the FCC [21]. ACTA claims that these companies should be regulated in the same manner as all long-distance providers and should not be allowed to "bypass" long-distance charges. Furthermore, ACTA claims that, unlike phone companies, providers of voice services over the Internet do not contribute to the Universal Service Fund and thus do not extend their service to the general population. ACTA argues that Internet voice services are "unfair competition," and it has asked the FCC to define what sorts of communications are permissible over the Internet. The ACTA petition directly targets two-way voice communications, but this development has direct implications for VC, which is also a two-way communication system. Suppliers of such systems may be susceptible to regulations applicable to telecom carriers. Microsoft and Netscape recently announced audio and video strategies for the Internet. Does this mean they are now telecom carriers? Regulating voice and video providers that make use of the Internet as a communications medium would appear to limit these providers' ability to develop alternative distribution channels.

The ACTA development comes shortly after the passage of the 1996 Telecommunications Act, signed by President Clinton on 8 February, which narrowly limits the FCC's authority to regulate the Internet [22]. However, a priority of FCC Chairman Reed Hundt is reform of the universal service regulations under the act [23]. Also, the act extends universal service obligations to all long-distance companies, not just the larger ones. The FCC is seeking comments on its public notice and is expected to make a judgment on whether to proceed with the ACTA petition by June. Complaints from phone companies about companies' use of the Internet for voice and video services are a new development in the United States, but the issue has already drawn the attention of governments in other countries, including Ireland, Switzerland, and South Africa [24].

4.2. Universal access

In addition to the ACTA petition asking the FCC to regulate what can be transmitted over the Internet, there is an ongoing public debate about whether the FCC should redefine universal service to include access to the Internet. This would involve the FCC in tracking areas of the country that do or do not have local dial-up access to the Internet and defining which services should be classified under universal access.

A two-year study by the Rand Institute concluded that universal Internet access for every American is "imperative" to the country's future and should be made a national goal [25]. The study advocates access to e-mail accounts and to appropriate computers and communications lines. However, various constituencies--private, nonprofit, and public entities--are concerned about what constitutes an agreed-upon definition of universal access. Does access refer only to an e-mail account, or will it also refer to other Internet features, such as the Web and audio and video communications? Access to a computer is implied, but to what type of connection? And is this access to be provided to each household (modeled after the telephone sector), or through a community point (modeled after the post office)? If the latter, public access points in libraries and other community centers would be established where the general public could gain access to the Internet.

Other concerns arise over who will pay for the information infrastructure and its use: Which mechanisms will cover the cost of providing effective access, and what are the public versus private roles in creating and operating such a service? Also, what role do individual states and communities have in implementing policy goals? The Rand study concludes that the free market alone is unlikely to provide universal access; it recommends government intervention to regulate the development of the information superhighway. Aside from the Universal Service Fund, we suggest an IP and domain name tax as one option to support the development of universal access through community points; another option is a tax levied on the general population.

5. Conclusion

We are witnessing the beginning of a significant phenomenon over the Internet: audio and video communications. The recent proliferation of suppliers of phone and video systems for use via the Internet does not necessarily mean that demand for such products will emerge. The silent underground movement of Internet phone users, however, gives us an indication that the demand for VC systems will indeed proliferate. The acceptance of the Internet as a new medium of communication has become the access avenue for VC to proliferate and become mainstream. This surely will not be a fad.

We mentioned options to price the Internet for congestion control and interconnection, and the question of whether the Internet is a threat to traditional telephone carriers. We briefly touched on recent policy issues, including the ACTA petition and universal access. We believe these must be taken seriously and are important to bring to the attention of the general population, where debate should take place over policy direction.

Acknowledgments

We thank Ke-Chiang Chu, Eric Emerson, and Eric Gradeler, whose comments have contributed to this paper. The views expressed in this paper are solely those of the authors.

Notes

[1] Hawthorne, N., The House of the Seven Gables and the Snow Image and Other Twice-Told Tales (Boston: Houghton, Mifflin, 1851).

[2] McLuhan, M., Understanding Media: The Extensions of Man (New York: New American Library, 1964). While McLuhan is often credited with coining the term "global village," Jarice Hanson, in Understanding Video: Applications, Impact, and Theory (London: Sage, 1987, p. 31), suggests the term actually originated with Samuel Morse, the inventor of the telegraph.

[3] Bell, D., The Coming of the Post-Industrial Society: A Venture in Social Forecasting (Harmondsworth: Penguin, Peregrine Books, 1976); Martin, J., The Wired Society (Englewood Cliffs, New Jersey: Prentice-Hall, 1978); Poster, M. The Mode of Poststructuralism and Social Context (Cambridge: Polity, 1990); Webster, F., Theories of the Information Society (London: Routledge, 1995, p. 1).

[4] Tang, J. C., and E. A. Isaacs, "Why Do Users Like Video? Studies of Multimedia-Supported Collaboration," Computer Supported Cooperative Work 1:163, 1993.

[5] Johansen, R., K. J. Hansell, and D. Green, "Growth in Teleconferencing: Looking Beyond the Rhetoric of Readiness," Telecommunications Policy 5(4), December 1981.

[6] For a thorough history of the Internet, see Hart, Jeffrey A., Robert R. Reed, and François Bar, "The Building of the Internet: Implications for the Future of Broadband Networks," Telecommunications Policy, November 1992.

[7] Hapgood, F., "I-Phone," Wired 3(10):140, October 1995.

[8] Bulkeley, W., "Picture-Phone Marketers Target the Home PC," Wall Street Journal, 27 February 1996, p. B1.

[9] This observation was made by McKnight in reference to high-definition TV systems, although it appears to be applicable to VC systems in our case. Refer to McKnight, L., "Economics of Interoperability," Internet Economics Workshop, MIT, Cambridge, Massachusetts, 9-10 March 1995, http://rpcp.mit.edu/workshops/cfp.html.

[10] Verity, John W., "Try Beating These Long-Distance Rates," Business Week, 22 April 1996, p. 131.

[11] Mills, M., "Freebie Heebie-Jeebies: New Long Distance Calling via the Internet Scares Small Phone Firms," Washington Post, 8 March 1996, p. F1.

[12] Consummate Winsock Apps List, http://www.cwsapps.com.

[13] The Internet may be prone to abusive behavior by its users, leading to a degradation in that resource, as in Hardin's "tragedy of the commons" (Science, 1968). Economic theory suggests that the Internet can also be included as a common good, since it is a shared resource common to all Internet users. This has led to the conclusion that the growth of the Internet generates positive network externalities gained through the Internet's reliance on interoperable protocols. Refer to Bailey and McKnight, who consider the Internet not only in terms of its economic feature of positive externalities, but also in terms of its technical characteristics of statistical sharing and its policy objective of interoperability. Bailey, J. P., and L. W. McKnight, "Internet Economics: What Happens When Constituencies Collide," Internet Economics Workshop, MIT, Cambridge, Massachusetts, 9-10 March 1995, http://rpcp.mit.edu/workshops/.

[14] The challenge, then, is to develop an effective Internet pricing mechanism but, in doing so, not lose the benefits of interoperability and the positive network externalities currently gained through the Internet's reliance on statistical sharing (Bailey and McKnight 1995).

[15] Bailey and McKnight 1995.

[16] Verity 1996, p. 132.

[17] Shenker, S., "Service Models and Pricing Policies for an Integrated Services Internet," in B. Kahin and J. Keller (Eds.), Public Access to the Internet (Cambridge, Massachusetts: MIT Press, 1995).

[18] National Science Foundation, press release, 14 March 1996 (contact: Mark Luker <mluker@nsf.gov>).

[19] Refer to Kahn, Alfred E., The Economics of Regulation, Vol. 1 (New York: Wiley, 1970) and Vol. 2 (1971); Panzar, John C., and Robert D. Willig, Contestable Markets and the Theory of Industry Structure (New York: Harcourt Brace Jovanovich, 1982); Fowler, Mark S., Albert Halprin, and James D. Schlichting, "Back to the Future: A Model of Telecommunications," Federal Communications Law Journal 38(2), August 1986.

[20] Federal Communications Commission. Report No. CC 96-10. Public Notice No. 2124: Common Carrier Bureau Clarifies and Extends Request for Comment on ACTA Petition Relating to "Internet Phone" Software and Hardware (RM No. 8775), March 1996, http://www.fcc.org/.

[21] Barmann, T., "Internet's 'free' Calling Irks Telephone Companies," Providence Journal Bulletin, 15 March 1996, p. 1F; Kramer, A., "Internet Calling Bypasses Old Phone Routes," Atlanta Journal and Constitution Daily On-line Guide, Cyberscene, 16 March 1996; "ACTA Coalition Attempts to Make Its Competition Illegal," NorthStar Newsletter, No. 13, 17 March 1996.

[22] http://www.fcc.gov/telecom.html; refer also to http://www.technologylaw.com/techlaw/act.html.

[23] The FCC was originally mandated under the 1934 Communications Act to provide universal service: the ability for individuals to purchase basic telephone services at a fair and reasonable rate regardless of geography. Extending basic telephone service to the general population was considered a public good that displayed positive externalities. See reference National Science Foundation (1996) for citations.

[24] Hapgood 1995, p. 140.

[25] Anderson, Robert H., Tora K. Bikson, Sally Ann Law, and Bridger M. Mitchell, "Universal Access to E-Mail: Feasibility and Societal Implications," Center for Information Revolution Analyses, Rand, 1995, http://www.rand.org/publications/MR/MR650/.