The 3G revolution, sparked by the merger of voice and data services in the cellular handset, has reached critical mass, say some industry watchers, while others see a slow ramp to the worldwide proliferation of third-generation handsets.
The turning point was reached on March 3, according to Ed Valdez, who directs platform marketing for Motorola Inc.'s Semiconductor Products Sector-a big supplier of ICs for the cell phone industry. That's when Hutchison 3G UK Ltd., a large U.K.-based customer of the Motorola division, officially launched its third-generation products and services-all implemented using Motorola's own i.300 platform.
Valdez pointed to the proliferation of wireless videophones demonstrated and used at the recent GSM World Congress in Cannes, France, as examples of the kinds of services Hutchison would provide.
But market researchers like Strategy Analytics (London) believe the worldwide shift to 3G is still on a slow ramp and 3G handset shipments won't amount to large volumes until 2005-06.
All of the standards needed for a solid conversion to 3G have not yet been fully adopted, explained Asif Anwar, a senior analyst specializing in RF components at Strategy Analytics. The conversion of Global System for Mobile Communications (GSM) phones into the Universal Mobile Telecommunications System (UMTS) will be crucial to 3G deployment, said Anwar. Wideband code-division multiple access (W-CDMA) would be the only means of ensuring the bandwidth needed for 3G services such as videoconferencing.
But rather than convert their 2G infrastructure to full W-CDMA, European service providers will deploy feature phones using the General Packet Radio Service (GPRS) modification to GSM, Anwar said. GPRS, the generalized packet radio service, superimposes Gaussian minimum-shift keying on the GSM waveform. This is largely a 2.5G deployment, he believes. "The baseband system suppliers will tell you they're working on 3G," Anwar told EE Times. "But what they're really selling is GPRS."
The highly touted deployment of versatile handsets in Japan initiated by NTT Docomo is really a limited phenomenon, Anwar believes. These are GSM phones with TDMA, TDC and PCS modulation schemes modified to accommodate pictures and game downloads. Downloads on the highly touted Matsushita P2101V video cell phone promoted by NTT, for example, produce only hazy video, according to reviewers. The explanation is that, without W-CMDA, 3G bandwidth is limited.
Meanwhile, Motorola's 3G reference platform, not unlike the hardware/software development platforms marketed by competitors Texas Instruments, Qualcomm and Analog Devices, includes practically all the hardware and software required to implement a portable phone that sports many features.
The company's reference platform includes digital signal processing and ARM core-based applications processor software for decoding MPEG-4 video, MP3 audio and JPEG still-camera images. It also includes radio modules based on silicon germanium and gallium phosphide heterojunction bipolar transistors for GSM, GPRS, Edge and W-CDMA-even Bluetooth and GPS, the worldwide global-positioning system. The cell phone maker can simply add its own distinctive packaging or user-interface software.
What will be added, however, remains a question. Though top-of-the-line 3G service packages, like those that include H.263 videoconferencing, are promoted in Europe, elsewhere opinions vary as to precisely what kinds of services will be implemented in the first and successive stages of 3G proliferation.
One reason for that uncertainty might be traced back to the standard itself. While the third-generation partnership project (3GPP) commercial standards document establishes a 384-kbit/second data rate as the recommended minimum for 3G, there is no universal agreement as to what services could be provided with that data rate. Internet downloads or GPS position finding could be accomplished at those transmission rates; but not Dick Tracey-style videoconferencing-at least not without significant compression and forward-error-correction coding.
Adding to the confusion is a footnote to the 3GPP standard that gave consideration to Qualcomm Corp.'s cdma2000-1x as a 3G technology. While the signaling provides only a 144-kbit/s data rate, it does act on the promise of combining voice and data through a cellular handset. This allows Qualcomm to promote what many regard as a 2.5 technology as 3G. "We're shipping 3G today," insisted Louis Pineda, Qualcomm's marketing vice president.
Certainly, Qualcomm's Brew (binary run-time environment for wireless) platform, which is used in phones manufactured by Kyocera and Samsung, is proliferating. But very few of the units being shipped to Korea and to many areas of the United States (via Verizon Wireless and Sprint) have streaming-video capability enabled.
There are roughly 25 million users of cdma2000, Qualcomm estimates. And while not ideally suited to video-streaming applications, the data rate is still substantially higher than that offered by a 56k modem, Pineda said. The company is developing enhancements to the technology that will provide data-optimized channels and transfer rates up to 2.4 Mbits/s.
Others, like Jim Koutras, worldwide marketing manager for 3G chip sets at Texas Instruments Inc., are also optimistic about the industry's ability to improve video-streaming capabilities. The DSP baseband developer says he was "pumped" after returning from the recent GSM World Congress in Europe. The quality of the video streaming he saw, even at 15 frames/s, ranged from "good to really good," he claimed.
Koutras believes that "2003 is when the pioneers like Hutchison and other green-field operators [service providers with a 3G license but no 2G foundation] will deploy. Anyone who moves ahead of the pack runs the danger of getting arrows in their back as competitors attack or attempt to improve upon what they've done," he said. Meanwhile, "The rest of the pack will deploy in 2004. "You should anticipate volume shipments toward the holiday season that year," Koutras said.
Many cell phone developers are attracted to TI's Omap development platform. Based on the ARM and TI DSP cores, the platform fits on a PC smaller than a business card and supports multimedia applications. Omap is also attracting attention among wireless PDA developers like HP's Jornada group; Sendo, a PDA cell phone developer (Birmingham, U.K.); and Palm's Tungsten group, Koutras said.
And 3G technology is already being deployed in basestations, Koutras said. Service providers need to test their systems and to work out the handoff and billing issues long before the services (and handsets) are offered to consumers.
U.K.-based Vodaphone is testing its 3G networks, as is Orange Telecom in France and Manx on the Isle of Man. "You want consistent quality if you're moving from cell to cell and having a videoconference-and no confusion as to how much it's going to cost you," Koutras said.
Indeed, "3G deployment is not a device issue-it's a service provider issue," said Chris Ambrosia, Strategy Analytics' director of wireless-device strategy services (Newton, Mass.). "Most service providers are cash-strapped, and do not have readily available funds to upgrade their networks," he said, so a move to video on UMTS will be gradual.
"Operators bid huge amounts of money for 3G licenses," said Doug Grant, Analog Devices Inc.'s wireless-business development manager. "With revenues in the tank, they don't have money to pay the government for those licenses-and no money to set up basestations. No wonder there's no handsets to turn on." Still, 3G is inevitable, Grant believes. "People want bandwidth."