There's nothing like information technology. Has any other industry sector transformed the world so rapidly and pervasively? It's unlikely, particularly considering wireless telephony and Internet connectivity. Growth in the number of people using these services in the decade straddling the new millenium is staggering.
From 1995 to 2005, the number of cell-phone subscribers rose from the low tens of millions, mostly in the U.S., to almost 2 billion?nearly a third of the world's population.1 The numbers of mobile phone users in Asia and Europe have each exceeded those in North America, with China being first among all nations and the U.S. now second.
Of course, population sizes matter. Penetration of the U.S. population totals about 70%. China, with five times the population, has less than 25%. But this bodes well for continued growth in the latter as well as in India, the other Asian mega-nation.
Internet accessibility has enjoyed equally remarkable growth in the past decade. It started with only a few million predominantly business and scientific users. Now, accessibility has become an almost ubiquitous feature of today's PC and probably its primary attraction for consumers. The number of people who have accessed the Internet worldwide is approaching 1 billion, though frequency and extent of usage are unspecified.2
North America again leads in penetration with 68%. Asian penetration is barely 9%, though it already exceeds our continent's user numbers and is growing at nearly double our rate. Numbers aside, the Internet's potential impact on society is much greater than that of mobile voice connectivity, though it has yet to achieve the average consumer's perception. Ultimately, much of this benefit will depend on ?broadband? connectivity, meaning high speeds and very low throughput delays, currently enjoyed by a minority of the worldwide users.
Given these dramatic growth statistics for wireless telephony and for broadband Web-based information and services, the popular view is that a marriage of the two inevitably will lead to even more rapid acceptance. A thoughtful analysis of current trends may lead to another conclusion.
Mobile telephony service for consumers has been available for about a quarter century. Analog telephony dominated North America until well into the 1990s, though it was practically non-existent elsewhere. Digital mobile telephony began in Europe in the early 1990s with GSM. By the end of the century, it had reached worldwide acceptance and predominance.
The European standards committee felt no need to be backward-compatible with a virtually inexistent analog technology. It selected a version of time-division multiple access (TDMA) somewhat inspired by the technology adopted by the communication satellite designs of the 1970s. On the other hand, North American carriers had no technology constraints for digital service. The Federal Communications Commission allotted each carrier 12.5 MHz but left them free to choose which digital technology to employ.
By the late 1980s, analog cellular already had several million subscribers. So the industry standards committee chose to be strictly backward-compatible with the frequency allocations, which were only 30 kHz wide and therefore quite narrowband. It nevertheless followed Europe's lead in selecting TDMA, but a TDMA that was severely bandwidth-constrained and weakened. This standard is designated by its standards document number as IS-54, but it will be called NA-TDMA here.
Another school of thought in North America embraced a truly broadband signaling approach. In place of the strict analog bandwidth constraints, it apportioned 1.25 MHz out of the total 12.5-MHz cellular band to a different digital technology. Code-division multiple access (CDMA) spreads the original narrowband digital signal over the full bandwidth.
CDMA is 40 times as wide as the analog and NA-TDMA signals, with appropriate digital signal processing. This provided relative immunity to multipath fading, the bane of mobile communications. Through interference suppression, it also could support many more users in a given spectral allocation?over an order of magnitude more than analog and about triple the sturdier version of TDMA, namely GSM.
This ?upstart? technology wasn't well received by either the European or American standards bodies and other critics. But after a number of successful demonstrations and numerous forums over three years, the North American standards body accepted the alternate CDMA-based standard alongside the earlier adopted NA-TDMA standard. This left the decision of which digital technology to use up to individual carriers.
A number of carriers chose CDMA, but the majority chose the NA-TDMA evolution. Analog was sufficiently entrenched in North America by the mid-1990s, so carriers were very slow in adopting the digital standards, unlike Europe and parts of Asia.
Surprisingly, South Korea embraced CDMA as its only digital standard and soon had tens of millions of subscribers. This gave CDMA its needed support to stimulate the manufacturing base. It also propelled Korean industry into the forefront of wireless manufacturers as well as service providers, helping turn the nation from an importer into a major exporter of electronics.
By the end of the century, there were nearly 100 million subscribers to CDMA service worldwide, with the bulk divided about equally between Korea and North America.3 Of course, GSM and its European base had expanded worldwide and reached nearly 500 million subscribers.4 Technology aside, GSM gained from three major advantages: a several year headstart, nearly worldwide roaming, and greater economies of scale. All digital technologies benefited equally from the relentless increases in silicon integration, as predicted by Moore's Law.
After about a decade of digital service and considerable consolidation of wireless service providers, CDMA users exceed NA-TDMA users in the U.S. by nearly two to one. And the gap is growing.
In the late 1990s, European regulators began planning for third-generation (3G) higher-speed data service. They also began considering spectrum allocations and technologies again. After some debate, they chose Wideband CDMA (W-CDMA). This derivative of the American CDMA differs primarily in bandwidth spreading to 5 MHz, rather than 1.25 MHz. It also differs in other parameters and some system features.
Additionally, the regulatory bodies in most EU countries chose to auction the 3G spectrum. Since it was the overenthusiastic late 1990s, the auction bids grew to unparalleled levels. The 3G licenses for the U.K., Germany, and Italy combined reached over 100 billion Euros. With this much invested, the incentive to field the infrastructure and produce handsets was considerable.
Nevertheless, the process took nearly five years to develop the technology. It was different and in some ways more challenging than the original CDMA, requiring backward compatibility with GSM, which is quite different. Meanwhile, the American, Korean, and Japanese carriers who had employed CDMA for their initial digital offering found the transition to the higher speeds of 3G much easier. This began five years ago in an embodiment called cdma2000.
It seems that the tables have turned. European carriers are struggling with the transition, while most American carriers are facing an easier technology evolution. With data speeds ranging from hundreds of kilobits/s up to a few megabits/s, 3G systems have gained momentum, reaching a total subscriber base of 50 million users with a predicted doubling within a year.
GSM dominates current digital technology, and W-CDMA has been made backward-compatible to it?hence the natural transition for GSM carriers. So, the number of W-CDMA users recently surpassed the subscriber base of high-speed cdma2000. The unanswered question is whether the EU carriers are now converting to 3G because of a perceived demand for high-speed data or to utilize the expensive newly acquired spectrum with the technology mandated by ETSI?and with a technology that's considerably more efficient than GSM.
Another wideband signaling technology, which originated in broadcasting and was adopted for wireline DSL and local-area wireless networks as well, has been proven effective in cellular networks through extensive trials. Orthogonal frequency-division multiple-access (OFDMA) spreads its signal by distributing it over many subcarriers, closely spaced in frequency and covering a wide spectral band like CDMA. This provides relative immunity to multipath fading and interference, gaining efficiency. Its other processing features show promise for greater gains too.
The Enigma Of Broadband wireless
The number of wireless telephone subscribers reached and exceeded the number of wireline subscribers in 2000, when there were 600 million of each. Today, wireless phones exceed wireline by a factor of three. So when will broadband wireless exceed broadband wireline? Before answering, we need to define ?broadband? and ?wireless.?
Broadband is relative and partially a misnomer. It really is used in place of ?high speed.? But of course, the wider the band utilized, the higher the speed supported. And in high-interference and fading environments, higher speed requires wider bandwidth. ?Broadband? applied to wireless refers to several hundred kilobits/s and above, but soon it will refer to at least a few megabits/s.
Wireless can have multiple meanings and refer to a variety of services and devices. Local-area networks (LANs) are mostly implemented by the IEEE 802.11 standards, commonly known as Wi-Fi. Personal-area networks (PANs) are implemented using the Bluetooth standard.
PANs typically cover less than 10 m, mostly to reduce unsightly and pervasive cables. Wi-Fi LANs with ranges of 100 m or more serve well within the home or the enterprise and typically replace the wired Ethernets. Their usage has greatly increased with the inclusion of Wi-Fi modems in many laptops and the establishment of Wi-Fi access points (basestations) in coffee shops, airports, hotels, and other public places.
PANs and LANs operate in unlicensed bands, usually around 2.4 GHz. With the exception of a few extended-range (a few kilometers) access points employing 802.11 technology, all WANs use licensed cellular bands. Wireless carrier decisions largely control the development and implementation of cellular wide-area networks (WANs). Yet manufacturers marketing directly to consumers have driven LAN and PAN developments.
LANs in particular have received a major boost by the inclusion of the modem capabilities in the PC microprocessors. Hence, they're practically offered free to consumers. Certain manufacturers are attempting to extend this success to WANs with IEEE 802.16e, which already is being marketed as WiMAX.
Whether it would be offered in a licensed or unlicensed spectrum is as yet unspecified. Nor are the costs and benefits compared to cellular well understood, especially given the massive expenditures in 3G infrastructure already invested by Verizon, Sprint, Vodafone, and other companies. Meanwhile, cities such as Philadelphia and Portland as well as the Bay Area see WiMAX as a means of providing broadband service to resdients who lack Internet access. Time and economics will determine the feasibility of this plan.
This also raises the question of whether such service needs to provide for terminal mobility or merely operate in fixed locations. Early versions of the 802.16 standard addressed fixed wireless service. However, 802.16e is specifically designed for mobile operation, which is more demanding and costly.
We need only consider WANs, since short-range wireless LANs are merely the data version of the cordless phone, a convenient extension within the home or office of the wired network. But we need to explore further the nature of the service and of the user's terminal. There are three overlapping levels of service?fixed, portable, and mobile?with increasing order of implementation complexity and cost and decreasing range.
Fixed service has the potential advantage of a high-gain antenna and hence greater range. But it can only be economically justified in locations where broadband DSL or cable is unavailable, unless governments subsidize the costs of infrastructure and service.
The argument for portable service is more compelling. Being able to access messages, information, and even entertainment wherever and whenever is indisputably valuable. Full mobility, referring to broadband usage while in motion at even high automotive or airplane speeds, may be less essential.
Terminals come in all sizes, from desktops to laptops to PDAs to ordinary cell phones, with practically a continuum of sizes and shapes. One accelerating trend is that the digital processing and storage capabilities require ever less size, weight, and power, so the terminal is becoming ever more portable.
Already, laptops have practically all the capabilities of desktops, and this trend is extending downward to PDAs. With the size of a notebook and the weight of at most a few pounds, terminals are becoming easily portable. Below the laptop, however, the limitation is the input-output, especially the screen size.
Portable broadband wireless service is becoming widely available in the U.S., Korea, and Japan. At least two major carriers in each of these countries have or are in the process of upgrading base stations in metropolitan areas nationwide to support broadband CDMA systems. The evolution to full mobility, which chiefly requires handoff among basestations, is also under way.
Portable service can be partially provided by paid subscription to Wi-Fi connectivity to ?hotspot? access points present in some coffee shops, airports, and hotels. The competition is through economics. Pricing in the U.S. for full wireless broadband service through cellular WANs is competitive with hotspot service and much more widely usable.
Service providers throughout the Americas, Asia, and Europe have installed infrastructure for broadband wireless data networks with basestation throughput speeds in the megabits/s. In North America and South Korea and partly in Japan, cdma2000 (EV-DO), the high-speed evolution of the earlier U.S.-developed CDMA (and hence backward-compatible with it), has predominated and gained about 20 million subscribers in the three years since its introduction.5
Following the lead of the largest Japanese carrier, European wireless carriers are implementing W-CDMA. This so-called 3GSM wideband service is backward-compatible to GSM to capitalize on the enormous (1.6 billion) user base of GSM. The relative immaturity of W-CDMA, coupled with the problem of backward compatibility to a disparate technology, has slowed its introduction until this year. Presently, though, manufacturers are predicting sales of 70 million such terminals by the end of 2005.6
Another driver for such handheld terminals is the fact that in Asia and even Europe, the penetration of the desktop or laptop PC is much smaller than in North America. Even with reduced capabilities, the handheld is the only device for Internet access for many users. There should be nearly 100 million termials enabled for broadband wireless by early next year, or about 5% of the worldwide wireless total.
The key question remains what these terminals and handsets will be used for. Judging from marketing strategy and cultural differences between continents, target user populations may differ widely. The first U.S. carrier to offer high-speed service aimed it directly at the business and professional user, offering a modem on a card for the laptop ?road warrior.?
Although a few PDA-size handhelds are now available with the same technology, they haven't been aggressively targeted to a large consumer market. Consequently, their price has been relatively high. Other applications that don't rely on high speed have been the consumer focus in North America.
The opposite approach seems to be under way in Europe, where 3GSM handsets are being offered at very low costs to ?seed? the market. The real issue isn't the technology used or its capabilities, but the applications users may want. And herein resides the enigma. Most cell-phone applications, even for the most advanced phones, haven't required or used broadband speeds. Some haven't even needed network connectivity.
The overwhelming majority of cell phones is used only for voice conversations. At about 8 kbits/s, compressed speech doesn't require high speed, though wideband signal transmission greatly improves quality and efficiency. E-mail messages require even less speed, though low latency is desirable.
Until recently, ringtones have been the most lucrative and most demanded feature. One-time downloading of games and other pastimes, also lucrative, haven't demanded high speed either. Web browsing and real-time video transmission require broadband, particularly in the forward direction. Two-way video conversations might further load broadband wireless, but it hasn't been particularly successful via any medium.
What are being heavily marketed are phones with built-in cameras and more recently high-quality audio, neither of which requires wireless connectivity. Position location is a space-borne wireless technology, but it's mostly independent of the cellular network other than for short control messages. A high-speed service soon to be offered will be television and radio broadcasts to cell phones or handheld terminals, but they will be transmitted on disparate spectra independent of the multiple-access network.
- GSM Association
- CDMA Development Group
- GSM Association
- airvananet.com: solutions - EV-DO Snapshot
- press.nokia.com: Feb. 14, 2005