Fixed wireless touted for broadband Net access
By Patrick Mannion
EE Times
(07/14/00, 4:55 p.m. EST)
eet.com
MANHASSET, N.Y. — Two new options for fixed wireless networks based on multipoint, multichannel distribution service (MMDS) debuted this past week, suggesting the technology may be ready to surge forward as an alternative to digital subscriber lines and cable modems for broadband Internet access. But disputes over modulation schemes, how to implement quality-of-service requirements and how to overcome the technology's nagging need for line-of-sight connections are far from over.
At the Wireless Communications Association International conference in New Orleans, six companies banded together to form the Wireless DSL Consortium. The group will tap the existing data-over-cable system-interface specification (Docsis) to deliver a version of MMDS that handles quality-of-service (QoS) requirements ultimately at a cost of less than $300 per subscriber unit. One member of the group, ADC Telecommunications, announced at the show that Austar, the largest pay TV operator in Australia, will deploy the technology to more than 60 cities in Australia, a new watermark for MMDS installations.
But the consortium quickly came under fire from startup Aperto Networks (Milpitas, Calif.), which rejected the Docsis approach as ineffective for handling the line-of-sight issues that dog fixed wireless. Aperto plans to use its own proprietary technology to roll MMDS products within 60 days, initially at costs of less than $1,000 per receiver.
Despite ongoing implementation battles, MMDS seems to be coming into its own as a bona fide high-speed, two-way data medium. Originally designed in the early 1990s for one-way wireless distribution of TV service, MMDS languished for years in the shadow of its cable and satellite counterparts.
The picture changed dramatically in September 1998 when the Federal Communications Commission blessed MMDS as a two-way medium, heating up the market.
Operating in the licensed 2.5-GHz region (3.5 GHz in Europe), MMDS is well-suited to transmit and receive high-speed data over distances of up to 35 miles. This alone has sparked the interest of carriers looking for scalable, low-cost, easily deployed technologies for high-speed Internet access.
To date, cable and DSL have not lived up to expectations in that regard. Even in cases where DSL has been successful, it is a victim of its own success, thanks to the crosstalk and signal degradation that occur as the number of active digital lines within a trunk rises. For its part, satellite data service still suffers from not having a return path other than the phone line.
That fact has made broadband wireless an attractive option for carriers. "We don't really care what medium we use, neither do our customers," said a spokesman for Worldcom (Washington). "It's not about fixed wireless [or any other medium], but about high-speed Internet access, and we'll use whatever technology gets it to the customer."
To date, Worldcom and Sprint have been the prime movers behind broadband wireless. Each has spent more than $1 billion to buy five MMDS companies apiece, in a head-to-head race to dominate the market on a city-by-city basis in the United States, Europe and South America.
Aiming to tap the hunger for alternatives in access networks, ADC, Conexant Systems, Gigabit Wireless, Intel, Nortel Networks and Vyyo formed the Wireless DSL Consortium this past week. The group leveraged the popularity of the name DSL even though its MMDS technology has no relationship to digital subscriber lines. (Several years ago, MMDS went by the similarly ironic moniker of "wireless cable.")
The rallying point for the consortium's effort is an air interface based on Docsis with enhancements to support wireless transport. Designated Docsis+, the specification is the result of work done by Vyyo Inc. (Cupertino, Calif.), a major vendor to system integrators and carriers such as ADC, for its trials in Mexico, and Worldcom, with its own trials in Boston.
Vyyo adapted Docsis for wireless by incorporating more-robust modulation schemes such as the quadrature amplitude modulation known as 16 QAM, along with quadrature phase-shift keying (QPSK). Docsis supports only 64 QAM. Vyyo also widened the frequency acquisition range to make it fit the wireless environment and handle frequency variations.
The new consortium believes Docsis will provide a proven route to supporting QoS as well as a relatively low-cost technology base. "Leveraging the volumes associated with Docsis in the cable market goes a long way toward lowering the overall cost of wireless broadband," said Arnon Kohavi, senior vice president of strategic relations at Vyyo.
Members expect to get the cost per subscriber device down to under $300 eventually. Right now, it hovers around $700 for most implementations. Said Worldcom's spokesman, "$300 is where it needs to be for the subscriber. To get it there, we may have to subsidize, we think."
As for quality-of-service, said Frank Kelly, director of marketing operations at ADC Telecommunications, "In a few months we're shifting to Docsis 1.1, which will add QoS provisions to the system." ADC uses Vyyo's equipment in its Axity Broadband Wireless Access System. "[Docsis] 1.1 hasn't been ported to the wireless environment yet, it's being done now," Kelly said.
"Without QoS, I honestly don't think a vendor can survive these days, and it doesn't matter what medium they're in," said Goli Ameri, president of research firm AmeriSearch (Portland, Ore.). "Bandwidth is becoming a commodity, service is the name of the game and how carriers can allocate their bandwidth into different services will be the differentiator."
ADC demonstrated voice capability at the show in New Orleans, but Kelly agreed that QoS is needed to guarantee voice bandwidth. "Other issues relating to network interfacing for voice must also be addressed," he said.
Using Docsis for MMDS has its opponents, including a separate consortium announced late last year and backed by heavyweights such as Cisco Systems and Motorola. This camp is using technology based on vectored orthogonal frequency-division multiplexing (VOFDM). But the most vocal early critic of the plan has been startup Aperto.
"The Docsis MAC [media-access control] assumes the channel characteristics are the same for every subscriber, and while that's true for cable, it's just not the case for wireless," said Alan Menezes, the vice president of marketing of Aperto. "As a result, the channels need to be individually tailored for each subscriber."
While Docsis proponents declare that upcoming modifications to the MAC and physical layers will take care of any outstanding issues over time, Menezes remains skeptical.
"They make compromises," he said. "With a Docsis MAC, when you design for LOS [line-of-sight], you don't reach all the users. When you design for NLOS [non-line-of-sight], you give suboptimum performance to those in LOS. You need to design for all these conditions simultaneously."
Handling non-line-of-sight customers is one of the strengths of the Aperto approach. Using a feature the company calls OptimaLink, the Aperto product performs dynamic control of link parameters to optimize each subscriber's connection. Each link employs 10 multilayer link parameters on a burst-by-burst basis to maximize coverage and capacity while maintaining a high level of spectral efficiency.
At the basestation, an adaptive algorithm selects the best set of parameters from each of 10 factors (such as power allocation, antenna diversity, frame size and modulation) to optimize a given channel. The result of this ability to process across a wide range of variables is a robust link with interference immunity and the ability to operate in NLOS conditions.
In the area of QoS, Aperto has its own technology that it claims connects seamlessly with industry-standard Internet Protocol Differentiated Services and Multiprotocol Label Switching protocols through back-haul networks. It is based on three service disciplines: priority service and unsolicited grants for constant-bit-rate applications; wireless weighted fair queuing for committed information rate applications; and round-robin service for best-effort applications.
Line-of-sight concerns
A third leg of Aperto's platform is its next-generation time-division multiple-access protocol. Designed for flexible bandwidth allocation and high performance under conditions of severe degradation by controlling transmission parameters, it supports both frequency-diversity duplex and time-division duplex schemes, maximizing spectrum use under global market conditions.
Both Kelly and Kohavi of the Wireless DSL Consortium are confident about their technology's ability ultimately to handle non-line-of-sight users. "We're addressing NLOS now with smart antennas, and with smaller cells with more power. It's a matter of how much power you want to transmit," said Vyyo's Kohavi.
"Our [approach] is near-LOS," said ADC's Kelly. "If you can't get it at 64 QAM due to a tree or some other obstruction, then it goes to 16 QAM. With the three modulation schemes in the equipment we have, you can serve up to 90 percent of subscribers from a single supercell. For capacity, you'll add many cells within that supercell to capture the rest."
Kelly added that if the costs of the VOFDM scheme used by the Cisco-led consortium come down, ADC might adopt that technology to reach the last few percent of users in an area where line-of-sight is an issue. |
