Lessons Learned

Most technology stories have neither a solid beginning nor a happy ending. Their beginnings are blurs of years, usually long periods during which a technology-perhaps one that was unearthed unexpectedly-is under development but still obscure and virtually unknown. In fact, most technology stories have no endings at all because the technological realm is cyclical and swirling, and few things in it are ever final.

Even so, a survey of wireless data's development is no ordinary historical jaunt. A common technological chronicle would examine the technical components and market drivers affecting a technology and service category, and interpret how those factors contributed to the eventual acceptance or rejection of the subject. However, to study wireless data's evolution is to document a pattern of frustration, patience, failure, incremental progress, more frustration and, ultimately, modest success.

To look forward, however, is much more encouraging. Wireless data has traveled a difficult road during its short life, but the one ahead appears much smoother, at least for now. The industry's recent focus on third generation wireless network evolution is driven not by a requirement to improve voice transmission, but by the need for high-speed data transport capabilities.

So while the industry argues over the legitimacy of and reasons for wireless data's potential upswing-whether they are the explosion of digital wireless network platforms, the ubiquity and popularity of the Internet and its Internet protocol (IP) transport base, or maybe just pure third generation marketing hype-steadfast proponents of wireless data must be beginning to feel at least a small sense of vindication.

Deconstructing CDPD The concept of wireless data began with cellular digital packet data (CDPD), a technology that was born when wireless network operators finished rolling out their analog cellular networks and were beginning to dabble in digital platforms and data applications.

CDPD likely was being developed for some time before that, but its public birthday was April 22, 1992.

"It was really a huge announcement at the time," says Ira Brodsky, president of Datacomm Research and long-time wireless data analyst. "There was a lot of expectation that data was another big market segment for cellular. It was like the marriage of the computer and wireless industries."

But it was pretty much downhill from there. Wireless carriers envisioned CDPD as a wide area mobile data network that they could deploy as an overlay to existing analog systems. Backed by most of the large cellular network operators, the system specification was developed to establish a common standard that would take advantage of unused bandwidth in the cellular airlink.

"The original idea was that it would be a channel-hopping solution and completely transparent to existing cellular networks," Brodsky says. "It looked to me like most of the early deployments were operating on fixed channels."

Technological issues and lack of a fast nationwide rollout were only the beginning of CDPD's ongoing troubles, he says.

CDPD's earliest adoptions were in decidedly unglamorous vertical markets, an indication that the technology's early mobile computing promises were not being met. According to the technology's critics, that is because it doesn't provide the bandwidth to support those kinds of applications and-although few if any network operators will admit it-because the characteristics of the CDPD airlink don't match those of its analog voice counterpart.

"CDPD doesn't have very good building penetration compared with the voice channels it is overlaying, so you aren't getting the same coverage," Brodsky says.

Presumably for that reason, most of the cellular operators supporting CDPD have focused on vertical market niches where the technology is logical but not necessarily profitable, including law enforcement, fleet monitoring, dispatch and telemetry-based monitoring.

CDPD is defined by the Wireless Data Forum as a connectionless, multiprotocol network service that provides peer network wireless extension to the Internet. It is designed to operate as an extension of existing data networks, namely those based on the overwhelmingly popular and ubiquitous IP.

CDPD's greatest strength-and perhaps the reason the technology is still alive despite the adversity it has faced-is that it supports IP. Because of that, any applications originally developed for CDPD can be adapted to run on any IP-based network.

"It's been a long, hard pull with CDPD, but the beauty of the decision we made to go with IP has really borne out," said Kendra VanderMuelen, senior vice president and general manager of the wireless data division at AT&T Wireless Services. "From an applications perspective, it's a very smooth migration story. Applications we have today move smoothly into the future products, and as we move, we end up with faster speeds."

The digital revolution Time has not been kind to CDPD. While the data technology was being relegated to somewhat obscure, low-profile tasks, three different varieties of digital networks were on the rise and threatening to eclipse CDPD altogether.

By carriers' admission, the earliest digital networks by incumbent cellular and newly formed PCS operators focused on voice. Network deployment is a multimillion-dollar, incremental buildout game, and voice applications already had proved that they could pay the bills, so carriers naturally took that route. Despite its perceived limitations, CDPD boasts an undeniably widespread footprint.

"All the digital networks that go up initially are geared toward voice traffic, and that's their market for several years," says Shirley Eis, president of Software Corporation of America, a developer of wireless and wireline data applications. "There's always limited coverage, and that's still true now."

But as those networks-based on code division multiple access (CDMA), time division multiple access (TDMA) and GSM air interfaces-matured, enhanced services in the form of wireless data applications came into view.

Supporters of GSM technology argue that their platform is best suited for data transport simply because of its design. GSM is an end-to-end standard that developed from a blank sheet of paper. Data was included from the inception of the vision, not as an overlay, as with analog networks and their digital cousins, proponents say.

"The protocols are designed specifically for data, which means they're geared toward making transmissions more error-free," says Roger Vigilance, vice president of product development at Aerial Communications, a GSM supporter.

"The other digital groups are trying desperately to catch up," says Roy Gunter, vice president of marketing and business development at Pacific Bell Mobile Services and head of the data committee for the GSM Alliance, a band of North American operators that support GSM. "They did not take data into account when they designed the initial technology."

Proponents of CDMA and IS-136 TDMA surely would disagree with Gunter's assertion that GSM is far ahead of the other technologies in its data pursuits. On the CDMA front, cellular provider AirTouch Communications has already begun testing its CDMA networks to determine the technological and marketing feasibility of applications such as fast Internet access.

"We're giving customers the gift of time by giving them more capabilities to manage non-voice information," says Cliff Fitterer, managing director of wireless information services at AirTouch. "It's not difficult to imagine a lot of demand for mobile information access."

Like most other wireless operators pursuing incremental digital buildouts, AirTouch is planning to use circuit-switched data first and eventually evolve to a packet-based solution. Moving to packet will help operators improve transport by boosting bandwidth and efficiency, Fitterer says. "The data rates will be evolving along with those transport capabilities," he says.

Fellow CDMA supporter Bell Atlantic Mobile is also moving toward a wideband packet data solution, but unlike AirTouch, BAM has continued to operate CDPD networks since the technology's inception.

"We're looking at a logical evolution of CDPD into CDMA packet," says Dick Lynch, executive vice president and chief technical officer at BAM. "The big future in data is going to be packet."

Most GSM operators are also pursuing a circuit-to-packet migration path, Gunter says.

"That will give higher-speed packet data to any operator that chooses to offer it," he says. "Packet offers a more competitive solution for customers who want to send regular data and lots of it."

AT&T Wireless stands out in the technology debate because of its McCaw-era decision to deploy TDMA technology rather than CDMA. Most of the carrier's current data efforts revolve around its PocketNet service, an integrated cellular phone and Internet appliance. PocketNet leverages the CDPD network and specialized applications to provide wireless access to the Internet and corporate intranets via a special browser equipped with Unwired Planet's UP.Link software. Current efforts to improve the bandwidth capabilities of IS-136 should expand the carrier's applications suite.

"Today we don't have a large market in general-purpose wireless computing in a laptop," says AT&T Wireless' VanderMuelen. "That is currently constrained by throughput."

That statement, by a representative of one of the industry's long-time supporters of CDPD, encapsulates the overall limitations of the technology and the reasons why any and all of the digital network platforms likely will prove more logical for data transport in the long term.

"There's a market for CDPD, but the question is whether [short message service], CDMA data and GSM data will blow right by it," says Brodsky of Datacomm Research. "I think they will unless CDPD can define some applications categories that it just wouldn't be worth deploying digital for."

Onward and upward One of the biggest topics in the wireless industry this year is the concept of third generation networks, or 3G. In general, 3G evolution is based on the global IMT-2000 initiative to upgrade all wireless networks to a common overhead standard for the new millennium. But exactly how those networks are going to evolve is anyone's guess.

There is no question what is really driving the 3G hype and the early evidence of the 3G effort: wireless data. Carriers want next generation capabilities so that they can port data over their networks in a timely, bandwidth-efficient way. They see the demand already on the wireline side, and whether or not they admit it, that is the side wireless carriers are striving to mimic.

"The crossover is going to happen in bandwidth," says Eis of SCA. "Moving more data more quickly is going to get more users."

AT&T Wireless moved quickly toward the future when the Universal Wireless Consortium, the trade group that backs the IS-136 community, announced an incremental plan for IS-136 evolution. At the Cellular Telecommunications Industry Association's Wireless '98 show in March, UWC released its three-phase program to meet the IMT-2000 requirements (Figure 1).

"It's really CDPD put on top of the IS-136 airlink," Eis says. "As we get to IS-136+ and have a 60 kb/s service, some of those [mobile computing] markets will get much larger."

Aerial Communications currently offers short message service (SMS) and a mobile office service for fax and file transfer over its GSM networks, but Vigilance is confident that plans for the technology's evolution will soon allow much more than that.

"There's a good migration path for how we're going to get from 9.6 [kb/s] to 14.4, 28.8 and 56 kb/s," he says.

PrimeCo Personal Communications, one of the PCS providers that supports CDMA, has a surprisingly different take on wireless data from that of some of its cellular counterparts. The carrier's data capabilities currently focus on SMS and its PrimeCast offering, which is primarily information services but eventually will include transaction-based services such as reservations, says Tom Sayor, executive director of product management at PrimeCo. That's not much different from other offerings currently on the market, but the carrier's evolutionary outlook seems more conservative.

"The initial step for us is to move toward two-way SMS to allow for interactive and transaction-based services," Sayor says. "Eventually we're going to move into more robust data, but we're going to focus on utilizing the existing handset base."

The real question surrounding 3G is what it ultimately could bring in the way of wideband data applications. Most operators are like PrimeCo in their hesitancy to answer that question and their unwillingness to compare those potential capabilities with those of landline networks. The only general consensus is that the future of 3G wireless applications likely will be determined by the users.

"You have to get a small army of early adopters out there and let them find ways to justify the use of the technology and the market," says Brodsky of Datacomm Research. "Most markets succeed when users start inventing applications."

"It's a matter of functionality, and the application is more important than the technology," says John Hainaut, director of wireless data and paging at Ameritech Cellular Services. "It's a business approach, not a technological approach."

PBMS is also in the process of testing GSM data transport for applications such as corporate intranet access, e-mail and Internet access with GSM handsets as modems.

"The demand is going to come from both the corporate arena and the consumer arena," says PBMS' Gunter. "We want to understand what it takes to support the customers in the marketplace."

Supporters of different technologies tend to agree that only time and practice will flush out what applications will prove most acceptable to end users. But they differ on whether wireless data will ever rival landline transmissions in transport speed and quality. That could be a technological issue or simply a philosophical difference.

"You're never going to see the bandwidth of wireless keep up with the throughput of a landline connection," says Hainaut of Ameritech. "There are clearly going to be compromises with wireless."

Brodsky and indeed many other GSM supporters disagree, arguing that 3G is all about wireless carriers angling for wireline business.

"Over the long term, I see this as part of the general trend of moving traffic from wireline to wireless," Brodsky says. "With 3G, we're talking about wireless being not only as good as a dial-up modem but better."

"The concept is to allow the network to do everything the landline network can do," says Vigilance of Aerial. "It's all in the early stages. I don't think anyone to date knows what the demand will be."

One of the things that plagued CDPD from the start was that it didn't make sense to customers. Even the most techno-savvy portion of the business community was still fairly new to wireless voice technology, and the concept of using wireless to transport data traffic was even more foreign.

For the current and next generations of wireless data to prosper despite that history, wireless carriers must step outside the technology and look at what they're doing from a customer point of view.

"Somehow we're going to have to push this in a way that's simple to customers," says Vigilance. "We need to take the complexity out. That's not a technological challenge, it's a marketing challenge."