Network/Computing Technologies

The electronic marketplace is where information-age commodities (digital, knowledge-based products) are produced, exchanged and consumed. For the market to function, it requires substantial investments in market infrastructure that is capable of supporting transactions efficiently. The market infrastructure includes telecommunications networks (telephone, cable, wireless, etc.) that provide communication and delivery services. Another key component is that of interface devices and technologies such as computer hardware, software and digital appliances. Network and computer technologies in turn enable various EC applications and other support services such as electronic payment systems. In this sense, networks and computers are enabling technologies of electronic commerce.

Enabling technologies will change and improve: messages have been delivered by person, postal service, telephone, fax, and now by email. Transportation and product delivery systems have been innovated over time. Just as likely, today's Internet and computing platforms will change. But these changes will not affect the fundamental nature of the digital economy. The electronic commerce revolution happened on the Internet because of its advantages such as open, distributed networking. Although technologies will determine the boundary of what processes can be implemented at a given time, it is the process we find useful that will shape the future of these technologies and the economy.

Both telecommunications networks and computer technologies are heading toward the "convergence". In short, the convergence is a process that allows hitherto incompatible devices to talk to each other (interoperable), products to be interchangeable and processes to be integrated. The underlying impetus for convergence is the digital technologies or digitization; its effect is much wider than a simple network convergence we are aware of today.

Product Convergence Today, digital products include voice signals, TV programs, musical CDs, videos, books, magazines, news and all types of paper-based information, database, computer software and games. Room and access keys are digital, stored on a smart card; IDs and personal information are digital; and money is digital. All these digital products are essentially made up of ones and zeros (the on-and-off dual states of an electronic charge), capable of being transported via a telecommunications network.

Network (Infrastructure) Convergence Telephone, coaxial cable, broadcast, satellite and wireless networks are all capable of sending and receiving the same digital signals. When a telephone network is busy, one can re-direct the voice to cable or wireless networks.

Market Convergence Suddenly, several regulated monopolies are facing potential competition from those who used to be in different markets: telephone companies and cable TV operators are potential competitors. Internet service providers and TV broadcasters may be fighting for Internet access. The convergence in products and networks has resulted in breaking down old market boundaries, in addition to lowering geographical market boundaries. (Of course, this potential competition will become real only if these firms do indeed compete. Another type of convergence, i.e. mergers, or artificial market boundaries will sustain monopoly status.)

Process Convergence We use process convergence in the sense of "integration" where different processes (or phases in a value chain) are integrated into a seamless process. For example, online advertising gets an immediate feedback from its recipients, which is then used to modify production process (customization). In broadcasting, advertising is almost separated from market research or any other business processes.

The convergence brings about new opportunities as well as uncertainties. As products are digitized, they acquire new characteristics increasing their appeal. For example, a CD-ROM version of an encyclopedia provides search and link capabilities far exceeding the cross-indexing features provided by book versions. New products mean new uses, new customers and new ways of doing business. Many focus on the opportunity to expand their business, but the novelty also creates uncertainty.

For example, as the telecommunications infrastructure converges, traditional boundaries among telephone companies, cable operators and satellite operators become unclear. These companies are experimenting video-on-demand services, interactive television, cable modems, online shopping, video dialtone, etc., to gauge consumer response and the future profitability in their widening playing field. Not knowing consumer demand and competitors' strategies, however, they are hesitant to plunge into the unknown. On the other hand, Bill Gates of Microsoft, Craig McCaw, who founded McCaw Cellular Communications, and other investors are willing to take a risk in the future of converging infrastructure. Their enterprise is called Teledesic Corp., which will invest almost $10 billion to place 840 low earth orbiting satellites around the earth. The plan is to offer broadband connection, broadcasting, video conferencing and other telecommunications services worldwide through its satellite network. The project's possible payoffs may be as large as the size of necessary investments.

The ultimate goal of a global network like Teledesic is to make the globe into your home base. Currently, local ISPs provide subscribers with local access telephone numbers for dial-up. This means that you have to pay for long distance calls when you are checking your email from different cities.

Roaming services often offer Internet access at cheaper prices than long distance calls. A few service providers have point of presence (POP) in many cities. Smaller ISPs can join an alliance so as to cover a larger number of POPs worldwide.

For example, iPass Alliance covers over 1,000 POPs in 150 countries. Any ISP allied with iPass will offer its own pricing schedule. (E.g. HomeGate). AimQuest is another collection of about 80 ISPs including Netcom which has its own 1,500 POP in 70 countries. Find out whether your ISP offers roaming service and if so get 800 numbers before you head out. Also in the work is Cencert Internet Dial Roaming Service, a joint venture by MCI, BT and Japan's NTT Data Corp.

The way we (or our computers) connect to the global network varies greatly. At work, computers are commonly "hard-wired" maintaining constant connection (i.e., through local area networks using Ethernet). Dial-ups establish temporary connections, typically through modems and via some telecommunications networks. These include the plain old telephone system, coaxial cables used for cable TV, wireless and cellular networks and satellite networks. Any of these connections may be constant if the line is dedicated to that purpose.

The physical link from your home to trunk lines (where the traffic coming out of individual users is aggregated and carried by high-capacity coaxial and fiber optic cables) is referred to as the last mile. The sheer number of individual users and the complex switching equipment required to manage traffic make the last mile relatively costly (representing between a quarter and a third of the total networking costs). At the same time, various telecommunications players are converging into this market. This will intensify competition among last-mile players. Will this competition lead to cheaper and better networking options for consumers?

Wireless Service Providers

Cellular: Cellular Technology Industry Association Richochet: Metricom Inc. Packet Data: Ram Mobile Data USA, and The ARDIS Company CDPD (cellular digital packet data): many sellular carriers offer CDPD, see CDPD Forum Smart Phones: AT&Wireless PocketNet, Wireless Data Division; and Bell Atlantic NYNEX Mobile (Cellscape) One-way Paging by most paging company Two-way Paging: SkyTel Corp.

Cable television network is built to handle one-way traffic: it is primarily geared toward delivering video programming to subscribers. Also its switching capability is limited to opening and shutting connections at the point of subscriber's residences. However, its coaxial cables can transport much larger traffic than twisted copper wires used by telephone companies. Telephone networks are excellent in terms of switching capability and managing two-way traffic, but the bandwidth is limited. In addition, its circuit switching network will have to be upgraded to packet switching (with routers and splitting devices) if the Internet traffic is to move on its network.

As audio (phone and cellular), video (cable TV, satellite and over-the-air broadcasting) and data network providers recognize the converging markets, they will invest in making their networks capable of supporting new types of traffic. However, their current markets will influence their investment decisions. One of the ways to minimize investment risks is to make new uses compatible with old ones. For example, broadcast and cable TV operators will want to tie new investments to their old business (ergo WebTV). Telephone companies are preoccupied with "long distance" nature of the Internet connection and how this figures into their capital expenditure and profits. The stake is high. In the networked digital economy, those who provide network services will combine the roles of telecommunications and transportation industries of the physical market. An ideal is to have an integrated or interoperable communications device for the last mile, but existing capital stocks and different interests of various telecommunications companies will present a challenge.

Telecommunication Information on the Web

Many sites related to the Internet infrastructure are listed in Telecom Information Resources on the Internet, maintained by J. MacKie-Mason.

Pricing network usage has become a hotly contended issue and the subject of growing field of Internet economics. The Internet economics, like transportation or telecommunications economics, analyzes the problem of resource allocation, in its case the limited resource of the networked Internet. Economists are primarily concerned with maximizing the utility and efficiency of a given resource. Prices are often found to be an effective tool in efficiently allocating resource among various users. Unless all Internet users put the same priority on their uses, prices will not be uniform.

Consumers are familiar with differentiated prices: telephone calls cost differently depending on time of use; cable subscription fees vary depending of the type and number of channels subscribed. Then, why is there such a great antagonism against usage-based pricing for the Internet? In part, this is due to the way the Internet was priced previously (i.e., flat-rate or no fees), and still substantial portion of the Internet users do not pay for their connection time. Furthermore, the mixture of paying and non-paying connection to the Internet brings about the problem of charging and paying for a particular usage. For example, some Internet users are forced to pay to receive unwanted messages. The lack of appropriate pricing and payment mechanism is one of the fundamental inadequacies of the Internet, not only as a market infrastructure but as an infrastructure of the future digital economy.

There are losers and winners for both flat-rate and differentiated pricing schemes. With flat-rate pricing, those with large consumption pay less than what they use, while those with small and occasional needs pay more than they use. In addition, flat-rate pricing tends to increase the overall level of consumption, increasing congestion and harming all users as a result. Finally, some with smallest demand may not be served (e.g. those who are not willing to pay the fixed monthly charge will not subscribe.)

Usage-based pricing may be able to levy fees based on what people use, in an equitable way if the differentiation scheme is equitable. The resource will be allocated efficiently in the sense that those who want to use will not be denied and those who doesn't need to (or unwilling to pay for) will not use the resource. The solution is efficient within the context of existing network infrastructure. But, users with high demand may prefer a flat-fee schedule for the obvious reason that it is cheaper. Also, prices are determined by users' willingness to pay and, if competition is low, prices will reach a monopoly level and there may be excess capacity even when some users cannot use the network. Whether that is acceptable is a policy question.

Network Pricing, Profits and Performance Usage-based pricing lowers congestion, but does it also generate more profits and/or customer benefits? A simulation experiment conducted at the CREC indicates that optimal usage pricing produces more profit, more consumer benefits and less congestion than AOL-type fixed fee schemes. See: Congestion Pricing More Profitable than AOL Pricing.

Finite Billing Unlike the case of telephone monopoly, the Internet traffic can be finitely differentiated and charged. For example, with telephone, the distinguishing aspects are time of use, location (residential or business) and the distance (local or long distance). But there is no way of knowing whether you are making a social call to a friend or transacting an important business deal. On the Internet, messages may be required to carry tags identifying not only the sender and the receiver but also the type of content. Based on the content, network carriers may be able to charge different prices: e.g., $1 for social email, $10 for stock information. (In effect, each of your phone calls are priced differently based on what you talk about on the phone.) While it is tempting to apply what we have learned in transportation and telecommunications networks to the Internet, it will not be a simple matter of just re-interpreting existing network economics and pricing for the Internet.

Related Web Resources

CREC Working Papers on Network Pricing Gupta, Stahl &Whinston. - An Economic Approach to Network Computing with Priority Classes Gupta, Stahl &Whinston. - Pricing of Services on The Internet Gupta, Stahl &Whinston. - Priority Pricing of Integrated Services Networks Gupta, Stahl &Whinston. - The Internet: A Future Tragedy of Commons? Gupta, Stahl, &Whinston - A Stochastic Equilibrium Model of Internet Pricing - PostScript version P. Konana, A. Gupta, D. Stahl and A.B. Whinston - Pricing of information services using real-time databases: A framework for integrating user preferences and real-time workload. In Proceedings of the 17th International Conference on Information Systems, Cleveland, OH 1996. P. Konana, A. Gupta and A.B. Whinston - Research issues in real-time databases in the context of electronic commerce- PostScript Version of Abstract Other Web Resources A list of links related to technical and economic resources about the Internet network infrastructure is the Network Economics site of the School of Information Management and Systems, UC-Berkeley. The Economics of Networks maintained by Nicholas Economides.

The end of NSFNET in 1995 signaled the end of research-oriented Internet. As the Internet has graduated into the commerce arena (despite the resulting problems), new initiatives are under way to launch yet another experiment for future networking technologies. Internet2 (I2) is a grass-root university test bed for advanced Internet applications, such as QoS (Quality of Service). I2 is built around ATM (asynchronous transfer mode), which is a standard for cell relay, a fast packet switching network. QoS refers to an IP network's ability to guarantee uninterrupted service when needed for high-bandwidth applications such as streaming video. I2 participants include over 100 U.S. colleges, each pledging to spend $500,000 as well as corporate partners such as Advanced Network & Services, Cisco, and IBM.

The Next Generation Internet (NGI) is the government's initiative. Clinton administration proposed $100 million per year for five years, but its funding will depend on the Congress.

Both projects have similar goals: to build a faster network for testing new technologies. Targeted speeds are 100 to 1,000 times greater than the current Internet (reaching up to 622 Mbps).

Related Web Resources

The NSFNET backbone that began in 1987 was retired in 1995 (the end of public Internet and the beginning of its commercialization). See a final report on the NSFNET: NSFNET: A Partnership for High-Speed Networking by K.D. Frazer. An interesting history of the NSFNET is given by S.R. Harris and E. Gerich. The proposed high-speed Internet backbone for the new Internet. Theoretically, regional networks are connected to the backbone at various network access points (NAPs) to this high-speed Internet. See, NAP maps.

Network Computers (NC) and NetPCs are cheap network devices that will allow more people to connect to the network. For electronic commerce to be ubiquitous, networking devices must be as affordable and accessible as television and telephones.

NetPCs are essentially stripped-down PCs priced lower than $1,000. With declining computer hardware prices and perceived consumer reluctance in buying machines with limited capabilities, the question is whether PC vendors will be able or willing to offer a reasonable machine at, say, $500. Its main backers are Microsoft, Intel and Compaq.

NetPC came as a response to Network Computers, which some predict will revolutionize the way we access networks and the Internet. An NC is a new type of an interface device to the networked world. It comes with TCP/IP, smart cards and Java but essentially nothing else. Software and contents reside in servers from which NCs will download whenever they are needed. Some equate NCs with dumb terminals of old connected to a mainframe computer. But an NC as Java Virtual Machine is world apart from a dumb terminal. A dumb terminal was simply a monitor and all computing was done at the server (mainframe computer). An NC will be connected to all servers in the network and have its own computing power. Oracle is the primary backer of the NC platform; Sun Microsystems provides Java software platform.

Skeptics argue that NCs are promoted by Larry Ellison whose main concern is to spite Microsoft and Bill Gates, and NCs will never compete with PCs. For one, PCs will be cheaper; for another, NCs are nothing more than dumb terminals. Other reasons? PCs are superior to any other computing platforms, and that people use their computers for games and entertainment for which a PC with full functionality is better. Microsoft and PC vendors certainly hope these are right. But none of these arguments are really sound. Do you remember those who laughed at GUI and mouse when Macintosh came out? They are the ones who now say Windows95 is the best thing that ever happened to computers.

The future of NC depends on how Java and applets develop and open up new types of services and uses of computer networks, including traditional communications. Smart card technologies will be an integral and important part of NC's success. Mac OS 8 introduced Java virtual machine in its operating system, with Applet Runner that can download and launch any Java applet without using a Web browser. Applets can be executable programs or plug-ins that can be used for a special task with existing software. Can you imagine how small and manageable Microsoft Word would be if it becomes a strip-down core program and all formats and tools are applets? Huge, inefficient and proprietary software is not the way of the future.

A PC belongs to an office or study; TV sits in the living room. A WebTV grew out of the vision that the mass market can only be developed if a device is not out of place in the living room. Does that mean a PC is for work and TV for entertainment? Can there be a hybrid of PC and TV?

Conceptually, a WebTV is that kind of hybrid. But if it aspires to be a PC like device to deliver the same old entertainment like a TV, WebTV will be a TV. Instead, it might prove to be an easy-to-use information appliance. But that will require new types of service, contents, delivery modes and usage so that knowledge-based products can be consumed in the living room.

WebTV therefore necessitates an integration of TV and PC, not just in technologies but in functions and contents. First, entertainment and information should be merged. Passive and active delivery must be enabled; it must offer two-way communication unlike TV or radio. Also, a hybrid of push and pull. Most importantly, WebTV or any information appliance must be able to function as an integrated communications, computer, and information device. Also, the ability to interface with smart products and household appliances.

In fact, we will not be watching WebTV, but we will be interacting with new information appliance in the living room, the one that will replace telephone, fax, TV, PC, cable, home security control panel, temperature control, mail box, etc.

Java is a programming language developed by Sun Microsystems. Applications written in Java can run on any computer. In addition to its multi-platform advantage, small programs (applets) can be written in Java and easily be downloaded whenever they are needed. JavaScript is a scripting language based on Java. JavaScript enables dynamic features on otherwise static Web pages. These functions are carried out by client computers (versus cgi scripts that run on the server computer), thereby distributing computing requirements.

Java is available for desktop machines as well as keyboard-less network device (WebTV), phones, pages, fax machines, and smart cards. Executable programs written in Java will enable the interoperable and distributed computing. Java's chief rival is ActiveX promoted by Microsoft. ActiveX (and its variations) aims to expand interoperability among programs running on desktop computers, an extension of Microsoft's Windows platform.

Java Resources on the Net

SUN Java site. Java FAQ list and tutorial . Belle Systems Java Resource Examples and links. Java Newsgroup JavaScript Page at Netscape

Think of an inline image displayed on a Web page. It is a digital file downloaded to your computer and your browser displays it using a helper program. An applet is a small program written in Java programming language. It is downloaded much like an inline image. Once downloaded to your computer, an applet runs a program on your computer, without a helper program.

Programmers can produce various applets to do nifty things, which can be run in any computer since Java is platform independent. Also, applets runs on client computers (i.e. yours), thereby reduce the work load of a server and minimize connection requirement.

Applets may change the way we exchange information. Today, information and knowledge are "written" in text files. In the future, knowledge can be made into executable programs: not only it tells you how to do something, but it actually does it!

Web Resources on Applet

Gamelan Java Applet Collection