Despite an ongoing technology war over 3G network standards, progress continues toward developing networks of the future

The main thing standing between code division multiple access and GSM networks of the future is a chip rate. But you'd never know it from the war of words that is erupting as a June 30 deadline approaches for standards bodies to submit their third generation radio transmission technology proposals to the International Telecommunication Union.

The standards groups peppered the media each week in May with news releases that insulted and accused one another of anti-competitive behavior. It takes a lot of digging to find actual progress, but beyond expelling energy to tarnish competitors' motives, supporters of the competing technologies are building a bridge toward 3G wireless networks.

Some players still hope for a single 3G network standard that operates worldwide in the 1900 MHz spectrum as designated by the ITU. At the least, most hope for interoperability between second generation and true 3G networks.

"Almost all have expressed a desire to find a way to harmonize the standards [so] that one can truly do things like global roaming," says Arun Sobti, vice president of research and development of standards and third generation cellular systems for Motorola. "People are looking at not just new applications but getting wireless ubiquitous, at a low cost."

The CDMA world wants to unify its wideband cdmaOne standard with the GSM camp's wideband CDMA.

But a stumbling block exists. The CDMA group wants the GSM camp to change its wideband CDMA chip rate to match that of wideband cdmaOne. GSM supporters say this would compromise their technology. The CDMA group says it won't and claims that the GSM crowd's balking will impede a single 3G standard's development.

Also, wideband cdmaOne synchronizes its base stations using global positioning satellites. Wideband CDMA does not rely on synchronization.

If the two standards remain separate, the U.S. ultimately may support three standards, including a time division multiple access solution called UWC-136 (Table 1).

This evolution of disparate networks is rooted in deep allegiances to particular technologies, with TDMA and GSM players hesitant to admit they likely will ultimately adopt a form of CDMA.

"Everyone is going to some form of CDMA. It seems this hasn't fully been recognized or digested," says Ira Brodsky, president of DataComm Research. "This is a very major transformation."

Religious wars aside, when will customers reap the benefits of these networks? Opinions differ.

"In conversation with Europeans, most are paying lip service to 3G because they want to get the spectrum," says Nicolas Kauser, chief technical officer for AT&T Wireless Services. European carriers don't operate in the 1900 MHz frequency, which is proposed for most 3G networks, so those countries will need to reallocate spectrum.

"But they aren't saying they will deploy it universally. They'll put it in where needed and then sell dual-mode phones," Kauser says. He points to countries such as Germany, where operators have just built second generation networks from scratch. "No one can afford yet to overlay a new network," he says.

Cost is not such an issue in other parts of the world where 3G networks are urgently needed.

"The Japanese will be first [to deploy 3G networks]. Whether the U.S. is right behind or not depends on the business case for 3G," says Gary K. Jones, director of standards policy for Omnipoint. If a particular application evolves that yields significant consumer demand, operators around the world will quickly follow the leaders.

"It will be faster than anyone is thinking here," Sobti says. "Once [the U.S.] decides to move, we move with amazing amounts of alacrity compared to others in the world."

Building a market Since the first mention of 3G, nearly everyone has asked two questions: What types of applications will make a 3G network pay off? And does a real demand exist for these applications?

Until recently, ideas for 3G network applications have centered on "multimedia"-the obscure and nearly sole descriptor of 3G applications used in past years.

"There is still some question about marketability and need for those kinds of services on a broad scope," says Ted Hoffman, vice president of technology development for Bell Atlantic Mobile.

Most players agree that the Internet will be the primary driver behind 3G. "There is no reason that people will only want to reach for a fixed location" to access the Internet, says Sobti.

Estimates put the number of mobile subscribers at 580 million and the number of Internet users at 500 million in 2001, says Brian Barry, president and CEO of Ericsson Canada. "Carriers want to carry that Internet traffic," he says.

Wireless data applications are expected to evolve to include on-demand video clips, personal automated teller machine banking, catalog shopping and work groups, says Craig Farrill, vice president of strategic technology for AirTouch.

Such advances challenge carriers to re-evaluate the context in which they consider 3G applications. Farrill describes the context as any combination of interaction among people, machines and networks. A machine could automatically query networks to offer information it knows a person wants. Farrill suggests that when PCs start to gather information in this way, they will compete with courier and overnight services.

Internet protocol (IP) is expected to play a major role in this wireless data progression. "The data world is going IP," Brodsky says. Among the applications he envisions are higher-speed IP telephony and high fidelity audio at high speeds using IP.

Despite these futuristic notions, there is still some concern over demand for them, especially considering data's slow growth history.

"The utility of high data rates is largely overstated," says Alan Sicher, manager of wireless data product management for Ericsson. "Data rate isn't so much the issue as is the ubiquity of coverage, availability of low-cost terminals, pricing and lastly, applications," he says. The industry must transcend the technology war and together work to create a market for data.

One hopeful sign in that direction is the change in name of the CDPD Forum to the Wireless Data Forum.

"The fundamental problem is creating awareness and promoting an end user value proposition," Sicher says. Promoting data in the next couple years will be crucial to the viability of 3G networks into the future, he says.

"There will be demand, but it has to be created," Brodsky agrees.

Predictions show that some demand should exist within the next five years. In 2003, 10 million of the estimated 900 million wireless subscribers will use multimedia, according to statistics cited by Mike Patriarche, vice president of strategy for Northern Telecom. These subscribers will account for about $11 billion in global equipment sales. Comparatively, existing networks and upgrades to them will account for nearly $40 billion in equipment sales worldwide, he says.

Much of that $40 billion likely will come from operators migrating to what is being called 2.5 generation networks-those with very high data rate capabilities. The 2.5 generation networks will be key for a few reasons (Figure 1).

"In the near term, operators will use their existing networks to provide those things that people really want now to cut cycle time down," says Perry LaForge, executive director of the CDMA Development Group.

However, in addition to cutting down cycle time, upgrading current networks also allows carriers to learn about market needs and the demand for 3G. "In large part, much of the 3G discussion has been a technical one, separate from the business case," says Bob Sellinger, director of wireless architecture and performance for Lucent Technologies. "They aren't anxious to deploy 3G if they can't think of applications for it."

Ultimately, he believes the business case will be solved by trading off speed and coverage for cost. In other words, high-speed networks might first appear in pockets and at incrementally higher rates to keep initial costs down.

Once customers start to use data, carriers likely can make a profit from it. An average data call lasts three to six times as long as the average voice call, says Harri Pietila, general manager of technology and solutions for Ericsson Canada.

Key to 2.5 generation networks will be converting today's networks to a packet-based delivery system, which allows networks to communicate with IP-based networks and services.

"IP really changes the way vertical markets can be approached today," Sicher says. In the past, proprietary data networks that operated parallel to voice networks were expensive to use and complicated because applications had to be customized to the network.

IP eliminates the need to rewrite wired applications to wireless protocol making the transition to wireless data applications smooth, Sellinger says.

Packet-based networks will alleviate data services' traditionally high prices-one of the major stumbling blocks for data, Sellinger says. "With wireless data, you're competing not only with other wireless carriers, but [Internet service providers]," he says. Lucent's research shows that users are willing to pay no more than 15% to 20% above the cost of landline for the benefit of mobility. That fact should force wireless players to price services to compete with ISPs' flat rates.

The road to 2.5 Vendors are developing higher data rate capabilities expected to go commercial in the next couple of years. GSM players are developing general packet radio service (GPRS) and enhanced data rates for GSM (EDGE) technologies. Combining the packet data solution of GPRS and high delivery speeds of EDGE, GSM operators can begin to offer significantly higher rate data services.

Nokia offers a path for GSM operators that includes high-speed circuit-switched data (HSCSD) and GPRS. HSCSD enhances existing circuit-switched GSM data and aims to match the speed of wireline services for connectivity to ISDN and public circuit-switched networks.

Nokia expects HSCSD, which can offer up to 57.6 kb/s and is implemented via a software upgrade, to be available this year. GPRS can offer more than 100 kb/s and is more compatible with LANs, WANs and the Internet. Because GPRS is a packet solution and GSM networks are circuit-switched, significant network changes would be required (Figure 2).

Nokia details three phases in the launch of GPRS, stretching from 2000-when the first commercial networks will roll out-to 2002 or 2003, when GPRS will have paved the way for third generation networks. "GPRS can be used as a radar to find where potential EDGE or 3G rollout could be started," explains a Nokia white paper.

Nokia recently formed a research agreement with Optus Australia and the Center for Telecommunications Information Networking in Australia to further research GPRS technology.

Rather than GPRS, TDMA players will evolve to IS-136 packet data that can achieve 64 kb/s.

After migrating to packet data, the next step for GSM and TDMA players is EDGE, which is considered a 3G technology by operators that cannot acquire 3G spectrum. It requires 200 kHz of spectrum, which AT&T Wireless plans to reach by combining seven 30 kHz channels to offer as much as 384 kb/s.

"It's a relatively smooth transition," Kauser says. He expects to reach 384 kb/s capability in two steps. Specifications for the first step, which will achieve 45 kb/s, should be complete at the end of this year with rollout at the end of 1999 or beginning of 2000. By mid-1999, Kauser hopes EDGE specs will be complete, with rollout to follow in 2000. All GSM carriers, however, may not necessarily deploy EDGE. "If it makes sense as an interim step, it'll be deployed," says Omnipoint's Jones.

CDMA players say their migration to higher data rates will be simple. The deployment of 64 kb/s data rates, which Sellinger says can be thought of as phase one of 3G implementation, can be done with a simple software upgrade and may start rolling out next year. The CDMA Development Group is working on the next evolutionary phase-144 kb/s in a mobile environment.

Ideas for holding down costs of future networks are already being discussed. Lucent is developing some innovative approaches to CDMA network architecture for data solutions. One example is using asynchronous transfer mode rather than T-1 connections to link switches and base stations. By using ATM, the data transfer is packet from end to end. Particularly in large networks, this type of network architecture can mean significant savings, Sellinger says.

AirTouch's Farrill points to other innovations that ultimately will help drive down costs for operators to implement what he calls "next generation" networks, including companies like Qwest that can help operators cut costs on long-distance transport.

Farrill believes that high costs of wireless radios may delay the point that allows 3G networks to be built. "Can we get a bit through the air at one-tenth of what it costs today?" he asks. That is the goal before 3G networks can become widespread-a goal he calls bold but possible.

Sprint PCS is the first U.S. carrier to announce plans for a trial third generation system. Working with Motorola, Northern Telecom, Lucent Technologies and Qualcomm, the carrier plans to test a 3G system no later than 2000.

The company stands uniquely aloof from the standards battle rocking the rest of the industry. "We've always said that our activities with 3G are not necessarily tied to industry standards," says Keith Paglusch, senior vice president of technical services and network operations. "If they work for us, great." If standards don't make sense to Sprint PCS and its existing infrastructure investments, company officials are confident that vendors will respond to its needs.

Sprint PCS has some distinct goals set for the trial. First, the carrier aims to ensure that 3G benefits customers. The new networks should improve service, provide more features and functions and offer faster services with more clarity, for example.

The new network also must be backward-compatible. "If we introduce a new network, we don't do it for the sake of new technology. We do it to be more efficient," Paglusch says. Sprint PCS supports wideband cdmaOne but continues to look into issues it might want to develop differently.

Until the trial happens, Sprint PCS is rolling out and developing different types of data capabilities, Paglusch says. "Once we do that, we'll have a good feel for additional data requirements."

AT&T Wireless is pursuing two paths toward third generation networks, depending on how the U.S. handles spectrum allocation. "We are taking a dual approach," says Nicolas Kauser, chief technical officer for AT&T Wireless Services. If the U.S. decides to clear and distribute new spectrum, AT&T will support wideband CDMA. "If there is new spectrum, we would love to have it to deploy a 3G network and support the world standard," Kauser says. "Even if the government made it available, though, it's not obvious who's going to get it." He says the likelihood of obtaining new spectrum in the U.S. is remote.

In the meantime, before spectrum decisions are made, AT&T is moving ahead with migrating its existing network toward 3G capabilities. Kauser believes that time division multiple access players occupy a strong position to migrate to the future because their current networks offer some 3G requirements. The main requirements for 3G networks are hierarchical cell structures and high-speed data. "We already have hierarchical cell structures," Kauser notes. AT&T is one of the few U.S. carriers that offers private, semi-private and public networks for applications such as wireless office services.

With the hierarchical requirement fulfilled, AT&T is now focusing on increasing its data speeds. The Universal Wireless Communications Consortium has proposed a migration path to higher speed data, with the first step allowing for 64 kb/s. This evolution can be done with a software upgrade, Kauser says.

As part of this evolution and the migration toward even higher data rates, TDMA networks will begin to converge with GSM. AT&T is working with GSM operators around the world to develop enhanced data rates for GSM. "Essentially what we're saying is we're getting convergence from IS-136 and GSM to provide data speeds," Kauser says.

Working closely with GSM carriers may be one of two choices TDMA players have today, says Ira Brodsky, president of Datacomm Research. They could develop their own technology that is completely different from the others and with little support across the world. "Or the next best strategy, which is basically an exit strategy-converge with GSM."