"This is good news for the industry and will really help to push the technology forward," Andrew Kreig, president of the Wireless Communications Association International (WCAI), said about the Intel announcement. "We needed to break the cycle of high cost which leads to limited deployment, and this will go a long way toward accomplishing that."

Announced in a keynote session at the WCAI conference here, the news from Intel follows a similar chip-development promise from Fujitsu in early April. The goal for both companies is to lower the cost of fixed-wireless-access (FWA) equipment enough to enable its mass deployment. The technology aims to supplant cable and digital subscriber lines in delivering broadband connectivity the "last mile" to homes and businesses.

Ever since Wi-Fi hotspots began cropping up across the United States, techies have dreamed of a net that melds wireless personal-, local- and metropolitan-area networks, allowing users to connect anytime, anywhere. But until recently the key piece of that vision, the wireless metro backbone, was missing.

A lack of standards and unreliable coverage contrived with the high cost of silicon to hamper mass adoption of FWA, relegating it to two niches: rural areas of developed nations, where cable and DSL have not penetrated, and undeveloped countries with little to no infrastructure on which to deliver voice or data. FWA has had particular success in South America, for example.

In much of Europe, where cable does not dominate and the phone system is too old to support high-speed data via DSL, interest in FWA has been high. British Telecom, especially, has increased its stake in FWA technology to further its promise of providing nationwide broadband access. But Europe's slow uptake of high-speed access in general has stalled adoption. Regulatory issues concerning the use of spectrum in the 2.8-GHz and 3.4-GHz regions have led to major delays, and in the U.K. and elsewhere, several innovative companies developing FWA equipment have gone belly-up or been forced to retrench.

But in the last six months, FWA's star has been rising. The IEEE in January finalized the 802.16a standard for FWA in the 2- to 11-GHz bands, complete with advanced quality of service and non-line-of-sight coverage. The standard encompasses single-carrier and multicarrier orthogonal frequency-division multiplexing (OFDM) modes with up to 256-QAM modulation, enabling bit rates as high as 74 Mbits/second in a 20-MHz channel over distances of up to 50 km.

Sriram Viswanathan, a director of Intel Capital's Broadband and Wireless Networking Investments group, endorsed 802.16 immediately, calling it "the next big thing" after IEEE 802.11-based wireless LANs.

Finalization of the standard was quickly followed by the formation of WiMAX, a Wi-Fi-like body with a mission to promote FWA and ensure the interoperability of equipment from member companies. The founders include Alvarion, Aperto, Ensemble, Fujitsu, Intel, Nokia, Proxim and WiLAN. On June 10, the group announced 18 new members, including Atheros and Andrew Corp., along with relatively small carriers such as SR Telecom and Towerstream. "The inflection point is happening now for FWA," said Reza Ahy, CEO and chairman of Aperto Networks (Milpitas, Calif.).

That may be so, but even with the presence of heavy-hitting silicon providers such as Fujitsu and now Intel, the success of 802.16a is far from assured. Greg Caltabiano, chief operating officer of Soma Networks (San Francisco), said .16a will never succeed as a point-to-multipoint solution for residential broadband access, although he did acknowledge FWA's benefit as a backhaul technology, which is Intel's primary focus anyway. Caltabiano's chief criticisms are cost and coverage, as well as the need to develop intellectual property from the ground up for the new standard.

The need for truck rolls, at $350 to $500 apiece, for each customer makes FWA a nonstarter, he said. "A Sprint representative was asked point blank at a WCAI conference panel this morning whether or not 802.16 could meet their broadband wireless deployment cost model, and he said no," Caltabiano said . "It doesn't matter if the silicon was free; the cost model just doesn't support carriers' needs."

Soma's strategy is to leverage the work already done on wideband CDMA and use that technology to provide downstream rates of up to 12 Mbits/s, with an upstream of roughly 2 Mbits/s, said Caltabiano. W-CDMA has proven coverage capabilities and no need for truck rolls, he said.

Another skeptic is Sai Subramanian, vice president of product management and strategic marketing at Navini Networks (Richardson, Texas). Subramanian cites everything from lack of major operator involvement and lack of real-world data from which to establish the specifications, to an almost zealotlike obsession on the part of 802.16 advocates with OFDM.

Instead of FWA, Navini and others — most notably, Flarion Technologies (Bedminster, N.J.) — are waiting for the proposed IEEE 802.20 air interface for mobile broadband wireless access. Still in the early-definition stage, that air interface aims for operation in the sub-3.5-GHz licensed bands and promises data rates up to 16 Mbits/s, with advanced mobility and handoff features and global compatibility.

Subramanian emphasized the major involvement of operators in 802.20, led by Sprint and signaled in a more circuitous manner by T-Mobile Venture Fund's investment in Flarion, announced last week.

"But that 802.20 is more a wideband, vs. a broadband, technology, so it won't offer the data rates needed for true high-speed connectivity for business-class service and Wi-Fi backhaul," said Ahy of Aperto, one of the WiMAX companies. Ahy added, however, that the .20 standard is "complementary" to .16a, echoing Intel's vision of a layered wireless-connectivity world. Those layers start with 802.15 and Bluetooth for personal-area networking, followed by 802.11 wireless LANs. On top of those comes 802.16a (and its mobile extensions, currently being defined under 802.16e) for metro-area networks and Wi-Fi backhaul, followed by 802.20 for wide-area networks to complement cellular.

It's the backhaul of Wi-Fi-based hotspots where Intel sees the potential for 802.16a, although its FWA chip sets will also be usable in customer premises equipment for the home and enterprise. While the .16a baseband is well defined, there is a lot of room for differentiation at the media-access control layer, said Ahy, and it is here that Intel will leverage the space-time coding and multiple-input, multiple-output algorithms it amassed in buying the assets of defunct FWA multiple-antenna specialist Iospan Wireless earlier this year.

Ronald Resnick, general manager of Intel's broadband wireless access division, said Intel will build the digital baseband and MAC layers and will work with radio-frequency specialists for the RF front end. The target spectrum is the unlicensed 2.45- and 5-GHz bands and the licensed 2.5-. 3.5 and 5-GHz bands. Alvarion (Tel Aviv, Israel) is working with Intel to define and test the chips on a system level, and expects to have product deliverable to customers by the second half of next year, said Rudy Leser, vice president of marketing at Alvarion.