Intel has called 802.16 "the most important thing since the Internet itself", and even allowing for a dose of self-serving, it is not talking entirely in hyperbole. WiMAX will extend the potential of Wi-Fi to far longer distances - around 30 miles. It is important, and is evolving at the speed of light (in standards terms anyway), but this rapid development is at the cost of the agenda being dominated by two already over-powerful vendors, Intel and Nokia.
WiMAX is showing off its first system profiles and interoperability tests at the the WCA annual conference in Washington DC this week, in a significant step towards making the 802.16a standard, ratified by the IEEE in March, a commercial technology. While a fully mobile version of WiMAX is in the wings, this first release will cover fixed wireless, and its supporters are focusing in particular on broadband last mile in unwired areas, and on backhaul for hotspots. Intel will start to make WMan chips this year and we should see WiMAX products early in 2004.
This is the opportunity for wireless technologies finally to grow up and offer the speed, multimedia support and ubiquity that Wi-Fi can never deliver. WiMAX president Margaret La Brecque is keen to stress the fact that Wi-Fi and WiMAX are complementary, but the newer standard holds all the real power. By providing a backbone for hotspots, based on standards rather than the various proprietary WLan expansion technologies out there, it makes the idea of a ubiquitous wireless network to rival cellular far more realistic than it ever was with Wi-Fi alone, despite the claims of the enthusiasts. The equipment makers are eyeing it keenly - amid all the doubts about the sustainability of the hotspot boom, anything that offers them a new product line plus helps to preserve the interest in Wi-Fi is to be welcomed.
For operators, there is a doubled edged sword. WiMAX operates in a mixture of licensed and unlicensed spectrum, and the initial products will be focused on 2.5GHz and 3.5GHz licensed and 5.8GHz unlicensed bands (though the full standard supports a far wider range of bands). The licensed spectrum gives operators the chance to apply for franchises for fixed wireless broadband provision, especially in rural and remote areas, something Intel is promoting assiduously as a means to increase investment in Centrino-enabled PCs (it now has a director of rural broadband access).
The unlicensed aspect means that independents have the chance to provide backhaul services for hotspots, which have the potential to create a nationwide wireless network. If the operators can control this, as they have been trying to do with Wi-Fi, they will be able to offer parallel, integrated services and achieve a stopgap as they struggle towards ubiquitous 3G - one with lower margins than cellular perhaps, but swifter ROI on lower upfront investment. They certainly have the power and resource to take control from alternative network suppliers, but they may also be condemning their 3G investments to stillbirth.
But the genie is out of the bottle now, and while the operators hesitate, the equipment makers are driving ahead, Intel in the vanguard, and Nokia, which has supported WiMAX from its earliest days, looking forward to the mobile standard and to the chance to add a new form of base station business to its ailing networks unit.
There are significant names missing from WiMAX so far - its initial focus on last mile is indicated by the bias of the membership towards fixed wireless, OFDM specialists (the 802.16 specification is built on an implementation of OFDM from Wi-Lan of Canada), rather than enterprise focused suppliers or mobile carriers. Some major vendors will be taking the usual gamble of trying to establish such market presence for their proprietary solutions as to sideline the industry standard - Motorola with its Canopy broadband fixed wireless platform springs to mind. But these companies will join - Cisco being a critical target - and in the meantime, the really impressive aspect of WiMAX has been its clear focus and unity of purpose. So far, perhaps because of its fairly low numbers, with most of these being smaller companies, it has avoided the complex politics and hidden agendas of most industry bodies - though this comes at the cost of a direction that is highly dominated by Intel and Nokia.
The best example of this is to contrast the political fiasco that scuppered the fast Wi-Fi standard, 802.11a, with the clarity of the 802.16a process. 802.11a had some technological shortcomings, notably its lack of backward compatibility with 802.11b, but its low uptake compared to the g' version was largely because it was continually delayed by wrangling between the IEEE and the European ETSI standards body. ETSI's HiperLan standard and 802.11a use the same frequency and so there was pressure to unify the two, but this was achieved painfully, with a separate implementation of "a" for Europe, which has further delayed product roll-out and confused users.
In contrast, WiMAX has set out from the start to harmonize 802.16 and the metropolitan area ETSI standard, HiperMan, and the specifications it showcases this week will demonstrate that unification. All this with remarkably little political in-fighting - the difference between leaving standards bodies to sort out their own futures, and putting a technology in the hands of vendors with a clear commercial objective, and deadline.
These vendors are finally giving broadband wireless the teeth it needs, with a standards base, to take on wired options for the last mile and for long distance networking. The WiMAX (Worldwide Interoperability for Microwave Access) group was actually set up two years ago by Nokia, Ensemble and the OFDM Forum, but gained a new lease of life in April when it was revived by Nokia in collaboration with Intel and added five new members, with nine more joining in May. The non-profit group takes a similar role to the Wi-Fi Alliance in WLans, backing development of wireless Man products based on 802.16 and working on standards certification and interoperability testing.
The initial version of the standard operates in the 10-66GHz frequency band and requires line of sight towers, but the 802.16a extension, ratified in March, uses the lower frequency of 2-11GHz, easing regulatory issues, and does not require line of sight. It boasts a 31 mile range compared to Wi-Fi's 200-300 yards, and 70Mbps data transfer rates.
Labrecque says that collaborating on mass market products will achieve similar economies of scale to those seen in Wi-Fi WLan devices. She says base stations will cost under $20,000 and support 60 enterprise customers with T1-class connections.
Systems based on the mobile version of the standard, which should ship towards the end of next year, about six months after fixed wireless products, will be able to achieve long distance wireless networking and will have far greater potential than Wi-Fi hotspots to provide ubiquitous coverage to rival that of the cellular network.
Although 802.11x cards will not work on an 802.16 network, the two could be bridged using a router between the two systems, effectively filling the gaps - and the inherent weaknesses - in the hotspot system. An extension of the standard, 802.16a, can operate over 50 kilometers and without line of sight connection, and could provide the last mile link between the national backbone and the high speed WLans. However, volume shipments are not expected for over a year.
Components of the standard
The WiMAX standard, 802.16a, was published in March. It extends the April 2002 line of sight fixed wireless Man standard, which focused on licensed 10-66GHz spectrum. The extension provides for non-line of sight access in low frequency 2-11GHz bands, some of which are unlicensed, and adds support for PMP and mesh technologies, plus boosting the maximum distance from 31 to 50 miles. Both of these standards went largely unnoticed until the WiMAX group was formed, led by Intel and Nokia, to promote and develop commercial solutions. Intel has now promised WiMAX versions of its Centrino chipset for next year, while Nokia says it will have battery and other technical issues solved in time to launch a WiMAX cellphone in two years' time.
The "b" extension is concerned with quality of service features, while the "c" extension focuses on interoperability. It relates to protocols, test suite structures and test purposes. Similarly, 802.16d focuses on fixing the errata and other protocols not covered by 802.11c.
The "e" taskgroup is working on enhancing the Wireless LAN air interface to support mobile as well as fixed broadband. It incorporates the work of the ad hoc committee, the Handoff Committee. A draft will be completed in July for ballot.
The "d" amendment creates system profiles for compliance testing for 802.16a devices. These tests will be in place within two months.
The overall vision for 802.16 is that carriers would set up base stations connected to a public network. Each base station would support hundreds of fixed subscriber stations, probably mounted on rooftops. The base stations would then use the standard's medium access control layer (MAC) -- a common interface that makes the networks interoperable -- to nearly instantaneously allocate uplink and downlink bandwidth to subscribers according to their needs.
Conceivably, 802.16 MANs could anchor 802.11 hotspots, which serve as wireless local area networks (LANs), as well as servicing end users directly.
The 10 to 66 GHz standard supports continuously varying traffic levels at many licensed frequencies (e.g., 10.5, 25, 26, 31, 38 and 39 GHz) for two-way communications. It enables interoperability among devices, so carriers can use products from multiple vendors and warrants the availability of lower cost equipment. The draft amendment for the 2 to 11 GHz region will support both unlicensed and licensed bands. µ
Sign up for INQbot – a weekly roundup of the best from the INQ