Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > RF/Microwave

WiMax upheaval on course

Posted: 25 May 2005 ?? ?Print Version ?Bookmark and Share

Keywords:wimax? broadband access technology? standard?

After two years of hype, WiMax broadband access technology is being realized in chips now and will be in systems by midyear. At that time, interoperability tests can begin in earnest, and the true nature and behavior of this so-called disruptive technology can be evaluated. The tests either will prove out the claims or expose yet another wireless technology's weak underbelly.

At face value, WiMax gives its proponents a lot to cheer: The standards-based technology and the ecosystem that supports it should yield far lower equipment costs than the expensive proprietary approaches that have historically-and spectacularly-failed to realize the promise of broadband wireless access. Operating in both licensed (3.5-GHz) and unlicensed (5-GHz) bands, WiMax features desirable non-line-of-sight capabilities, derived from its use of 256-tone orthogonal frequency division multiplexing (OFDM). "That's the main reason for deploying WiMax," said Carleton O'Neal, vice president of marketing at Alvarion Ltd.

Coupled with adaptive modulation and coding, up-channel subchannelization and beam forming, WiMax promises blanket coverage with distances as far as 30km and data rates up to 75Mbps (though typical expectations are for 2Mbps to 3Mbps at about 5 miles, depending on the environment). Throw in an all-Internet Protocol structure and advanced quality-of-service mechanisms, and WiMax inspires an enthusiastic following.

Led by Intel, those supporters are committed to leveraging WiMax's features to provide cellular backhaul, last-mile access for rural and urban consumer and enterprise subscribers, and even data offloads for 2.5G and 3G networks in densely populated areas. Most expect WiMax to have the greatest initial impact in areas of the world that lack a dense legacy infrastructure, such as Latin America, much of Asia, the Middle East, Africa and Eastern Europe, as well as rural regions in more developed countries, including the States.

"It's disruptive," said Mohammad Shakouri, vice president of marketing for the WiMax Forum, "but not because it'll replace 3G cellular-because it's based from the ground up on IP and an open network model. The IEEE and consumer companies are following a different model to the regulated ITU and telecom." Shakouri predicts an equipment market in the $1.5 billion to $2 billion range.

And all that's just for the fixed version, based on IEEE 802.16d-2004. Later this year, the IEEE expects to put the finishing touches on a mobile version, called IEEE 802.16e. A key part of that effort is harmonization with the South Korean WiBro standard, which is set to launch in that country in the 2.3-GHz band. Once the WiMax Forum generates the profiles for the mobile version of WiMax, a move expected in early 2006, proponents hope to enable services that offer subscribers a full multimedia experience, with voice-over-IP (VoIP), video, data and audio, and with advanced hand-off capabilities to enable full mobility.

The mobile experience may be enhanced through efforts to open up the 700MHz band, which would allow for longer propagation and better through-wall penetration. Some question the business case for that, however, wondering about the difficulty of frequency reuse and when-or even if-spectrum will be opened up in that band. In addition, the spectrum that might become available is both narrow, at 6MHz, and scattered.

The 700MHz option is "interesting from an RF propagation standpoint, but how it gets allocated is an issue," said Derek Dicker, director of strategy planning for Intel's broadband wireless division.

That being said, in February, the Federal Communications Commission granted Aloha Partners a waiver that will allow it to begin tests in the 700MHz spectrum in Tucson, Arizona.

Regardless of what happens at 700MHz, mobility is still the endgame for WiMax, with datacom innovation thereafter expected to proceed at a hypermetabolic rate, starting in early 2007.

"We're taking Ethernet and making it mobile, noted Alvarion's O'Neal. "Can you think of a bigger opportunity?"

But not everyone's convinced. Nambi Seshadri, vice president and chief technology officer for the Mobile and Wireless group at Broadcom Corp. (Irvine, Calif.), sees High Speed Downlink Packet Access, the latest revision to the 3G W-CDMA standard, as "the path of least resistance toward true wireless wide-area broadband connectivity." The company prefers to focus on that path, though it's keeping a watchful eye on WiMax.

Texas Instruments, another company deeply committed to the cellular market, has also refrained from taking a formal stance on WiMax, though TI believes its technology and platforms are ready and able to support it.

Atmel Corp., which is developing WiMax radio front ends, sees a viable model for the technology in "data to the home at a reasonable price for urban, suburban and rural areas," said Chris Baumann, director of BiCMOS products.

And Bernard Aboussouan, VP of marketing and business development for startup Sequans Communications (Paris), is betting on fixed WiMax and its potential for nomadism and portability to generate initial revenue. But then it will play with WiBro, which is expected to launch in South Korea next year. In that way, if the mobile version of WiMax does not gel by year's end, WiBro would be a viable market.

But what WiMax proponents don't acknowledge is that "at 3.5GHz and 5.8GHz, you need an awful lot of basestations," said Ronny Haraldsvik, VP of global communications and marketing at Flarion Technologies Inc., a manufacturer of broadband wireless equipment based on WiMax competitor Flash-OFDM. "That's why they're working so hard to get the costs down and going closer to the cellular bands [700MHz]. But we've been focused on mobility from the start."

And there's still a question of timing. Though the standards and profiles are set, interoperability tests-initially with Cetecom Spain-won't start until July, which will push out final system availability until later in the year. Only at that point can carrier trials begin, and those trials can take up to 18 months, said Haraldsvik.

Flarion spent three years in trials with Nextel before the tests were dropped when Nextel merged with Sprint, which uses CDMA. According to Michele Pampin, a director of strategy at Harris Corp., carriers are rapidly deploying 3G basestations to meet data-rate and coverage demands and want a wireless backhaul solution to eliminate the multiple, expensive T1 lines. And they want it today, not a year or two from now.

"Backhaul is a huge concern for carriers," said Pampin, who sees double-digit growth in the number of basestations being deployed, some with up to 12 T1 lines. "It's a global phenomenon."

As a result, carriers are forced to opt for proprietary solutions until such time as WiMax systems become available. To meet that interim demand, Harris will use this week's CTIA conference in New Orleans to launch its TRuepoint 5000 microwave point-to-point system. The company recently announced that Sprint will deploy TRuepoint systems as part of a four-year agreement to provide backhaul for PCS systems in key U.S. markets. That said, both Sprint and Harris are active within the WiMax Forum; Harris is a founding member.

Finally, there's the issue of whether to deploy fixed WiMax (802.16d) now or skip it altogether and wait for the mobile version. "With Korea launching [WiBro] next year, many see [802.16]e as leapfrogging [802.16]d and even 3G," said Rupert Baines, vice president of marketing at picoChip Designs Ltd, a developer of baseband processors for WiMax. An aggressive push by Samsung and recent comments by the Indian and Japanese governments in favor of WiBro add to the concern.

But Jeff Leasure, SiGe BiCMOS product manager for Atmel, believes the timeline for getting 802.16e finalized will force many outside South Korea to deploy fixed WiMax first. Leasure, who is involved in a number of WiMax Forum workgroups, said the timeline the forum has laid out calls for 802.16e equipment to emerge in early 2007.

That may be optimistic, however, given the current turmoil within the 802.16e group as WiBro proponents debate with Intel and Qualcomm over harmonization and how mobility should be achieved. There are debates over whether the hybrid automatic repeat request (Harq) scheme being pushed by South Korea should be incorporated and how multiple-input, multiple-output (MIMO) smart antennas should be implemented. Intel is also pushing low-density parity coding.

"No one seems to be able to agree on how to do MIMO," said Ron Murias, manager of applied R&D at future WiMax systems vendor Wi-Lan Inc. (Calgary, Alberta), who believes Harq is premature. Uncertainty over patents and IP rights also persists.

The solution, according to Atmel's Baumann, is to deploy .16d now and then perform a board upgrade when the time comes. "The biggest change is in the baseband," he said. "You'll notice the basestations are not highly integrated, so you can swap out MAC/PHY boards."

Murias would go one step further toward simplicity. WiBro's 2,048-tone multiple-access scheme, compared with WiMax's 256 tones, "means this stuff [harmonization] is really complex," he said. The best solution, he believes, is to deploy OFDM 256 for both fixed and mobile WiMax. That will let carriers perform a software upgrade to their systems to achieve the MAC functions of 802.16e while avoiding the need to swap out the PHY board until the market is proven.

While skepticism about the next big thing is justified in the wireless world, the WiMax Forum's roster nonetheless reads like a who's who of semiconductor, board, software, system and telecom companies-and it continues to grow. As for product realization, Intel has announced its Rosedale 5116 client-side-only MAC/PHY chip and has lined up Aperto, Alvarion, Proxim and Airspan as system-level partners. Wavesat Inc. has been shipping its DM256 PHY-only chip, a sixth-generation design, since December. Targeting both basestations and client devices, the chip can be accompanied by Cheetah, an FPGA-based implementation of the 802.16d MAC. Fujitsu, meanwhile, will reveal the details of its MAC/PHY chip in late April at Broadband Wireless World in Las Vegas, said George Wu, director of marketing for the Technology Solutions/ASSP group of Fujitsu Microelectronics America Inc.

Sequans is in the FPGA stage with its MAC/PHY system-on-chip, but, according to Aboussouan, the company expects to have a final chip midyear for 802.16d. A WiBro-compliant mobile version is planned for year's end. Sequans recently announced a further $9 million in funding.

For its part, Atmel is pursuing the RF front end of the signal chain, believing both that such a focus is the best application of Atmel's expertise and that current WiMax radio designs are deficient, Baumann said. "We want to have a radio that's driven by the standard, not a modified wireless-LAN radio," he said.

The latter are doomed once interoperability tests start, he said, since many use direct-conversion architectures and thus won't be able to meet the phase noise and timing requirements of WiMax.

Whatever the architecture, flexibility is the key, said Baines, particularly in the face of interoperability tests. That's the foundation of picoChip's PC8520-16d platform.

'Programmability and flexibility'
Jules Pierre Lamoureaux, Wavesat's vice president of engineering and chief technical officer, couldn't agree more, having been stung by opting for an SoC out of the gate. "Our next generation will be an SoC for the customer-premises equipment and an FPGA for the basestation, as we need the programmability and flexibility for upgrades," he said.

That philosophy is also behind Analog Devices Inc.'s decision to partner with SiWorks in a deal that will see SiWorks develop a basestation PHY layer on ADI's TigerSharc general-purpose DSP. "WiMax is very new, with lots of options and a migration path to 802.16e," said Jeff Stevens, manager of the broadband wireless group for ADI. "It needs to be flexible."

Other considerations include the degree to which, and how well, WiMax silicon and system vendors have implemented the various options. This is particularly the case for beam forming and subchannelization, Baines said. Non-line-of-sight effectiveness could also be a deal breaker.

- Patrick Mannion
EE Times

Article Comments - WiMax upheaval on course
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top