Infophones to bear heavy loads

Infophones to bear heavy loads
The problems of designing a really good third-generation cellular phone are strangely similar to the hurdles engineers faced defining pen-based computers in the late 1980s or building interactive TV boxes in the early '90s.

As with these overhyped systems of the past, no one knows exactly what applications third-generation phones are going to run. And expectations are once again schizoid: The devices are expected to do quite a lot and cost very little.

But perhaps what makes the 3G phone conundrum especially knotty, albeit exciting, is the amount of change going on at multiple levels with applications, networks and systems.

"You have three or four trains moving in parallel," said Rajiv Mehta, director of business development for Motorola's iDen phones, which are used on the Nextel network. "There's the migration to higher-speed networks, the rise of Java as an application environment and [the advent of] devices with larger displays and more processing and memory. The key is whether we can define the apps that tie all these things together."

There's no shortage of apps being mentioned as possibilities for tomorrow's phones to crunch. The laundry list includes e-mail, Web browsing and various flavors of text and multimedia messaging that might appeal to all comers. It also includes games, and taking or viewing color pictures or video geared for consumers; and access to corporate data and applications for business users; plus all manner of permutations of those apps. Others being dreamed up today that are more specific to the cellular world include location-based services for phones with global-positioning-satellite radios.

No one has plucked a killer app from the heap-some say there doesn't need to be one. But many OEMs are working overtime to parse the applications into segmented markets so they can design their phones for targeted groups of users. Nokia, for instance, shows a matrix of about six price points across about eight target markets. Some apps are particular to one category-like 3-D games for high-end consumer phones. Others, like messaging and information services of all stripes, span many segments though they are customized for, say, business vs. consumer users.

"We have a really good feeling about applications in a given segment," said Joe Coletta, director of applications management for Motorola's cellular phone group. "We are at different points of development with different apps. In the near term you can see the arcade games like PacMan coming up. Longer term, location-based services will be big."

Many observers say that three tiers of phones will emerge. High-end models will merge PDA and cellular functions, many using Palm, Symbian or Microsoft Smartphone operating systems. A broader middle tier of Java phones will handle many applications but fall short of a PDA-class experience. And a vast bottom tier of enhanced voice phones will offer messaging and games for a growing consumer population that increasingly includes kids.

One arcane acronym has arisen as a guiding principle out of the uncertainty of this applications soup: "Arpu," or average revenue per user. That's what accountants at network carriers are projecting on their calculators today as the return on investment for the billions spent obtaining spectrum and deploying new network systems. Data services represent only 7 percent of cellular revenue today, but Nokia projects that will grow to 35 percent of a growing pie by 2006.

The new phones must be able to dial carriers into these rising Arpu rates. They will do this by enticing users to buy new apps and subscribe to new services. "Carriers want to get to $2 to $3 a month in new subscriptions," said Peter Bernard, product-marketing manager for Insignia Solutions (Fremont, Calif.), a Java virtual machine provider that is relaunching itself with tools to help carriers update cell phone software.

It's worth noting that AT&T Wireless is borrowing a page from its Japanese partner NTT Docomo's book, with plans to charge flat transport fees and let service providers take the lion's share of revenue from buying new apps or services.

For the OEM, this all means two things: These phones will have to run a broad array of applications on a relatively new and unproven software stack deployed by a variety of companies. And that software stack will be a dynamic one, probably changing more often, with code from more companies, than today's often unstable desktop PCs, which run software primarily developed and deployed by just one company.

As if this were not a big enough design problem, the phones will have to be inexpensive. "I see the cell phone makers really struggling right now. As if the GSM vs. CDMA [network modulation] unrest wasn't enough, people now expect that cell phones are free, so the product has almost no value-but now it's got to do all these cool things," said Andrew Burt, a wireless business development manager for Toshiba. Burt has the challenging task of selling MPEG-4 silicon to cell phone makers interested in devices that could play video clips. His only design win at press time was with the cell phone division of Toshiba.

David Brudnicki, director of technology for a multimedia group at AT&T Wireless, reports the costs of the initial cell phones the carrier is deploying on its 2.5G networks rolling out this year are no higher than past generations of high-end phones. However, Brudnicki also noted that "phone subsidies are the bane of the carriers. How do you get to profitability if you have high acquisition costs for new subscribers?"

That suggests users could be asked to pick up more of the tab for 3G devices, on top of the new applications, services and transport fees they will be expected to pay.

Yet, game developers are psyched. Ian Baverstock, business development director for Kuju Entertainment, a developer based in England that has been writing games for cell phones for two years, estimates that market today to be less than $1 billion, 85 percent of it in Japan. But Europe and the United States could represent a $6 billion mobile-game market with 198 million players by 2005, he said. Baverstock estimates the world market could grow to as much as $12 billion by then.

Carriers are certainly moving ahead optimistically. AT&T Wireless was 40 percent finished with its 2.5G network rollout at the end of 2001 and expects to have a nationwide network done in the United States at the end of the year. It will then immediately begin deploying faster Edge technology for GSM networks. "This is a year of transition for developing and deploying our network," Brudnicki said.

Ditto Sprint, which is moving just as fast for its code division, multiple-access networks. The company is deploying capabilities for peak 144-kbit/second data rates nationwide by the end of this year, following up with peak 288-kbit/s technology in 2003 and promising 3- to 5-Mbit/s capabilities by the end of 2004.