Analysis Intel this week pledged to "solve in silicon" the incompatibilities between the mobile and fixed versions of WiMAX, the wireless broadband contender.
According to Sean Maloney, the head of Intel's Mobility Group, these two WiMAX modes are not the same.
How come? First, a recap. WiMAX is the brand name given to the 802.16 wireless broadband standard approved by the IEEE in 2002, and has updated almost annually since then. Revisions have eliminated the need for a line of sight, for example, and added support for more spectrum bands. The most recent version of the standard is 802.16-2004, approved last June, after the proposed 802.16d revision of the standard was formally accepted.
Whether its called 802.16d, 802.16-2004 or "wireless DSL", the standard details a wireless broadband connection, typically between a service provider's base-station and an antenna mounted on the customer's premises. In other words, a fixed link.
Intel is keen on this because the more punters have access to broadband Internet connections, the more online content services they will want to use and - it believes - the more powerful computers they will desire. So it hopes WiMAX will enable it to sell higher-margin microprocessors and chipsets as well as wireless chips. This is doubly important because it expects wireless chips to become a commodity relatively quickly.
Intel is already shipping 802.16-2004 silicon, with a view to its customers offering WiMAX kit for sale in the second half of the year.
The chip giant also wants to see WiMAX integrated into notebooks the better to enable wireless Internet access with all that tedious mucking about in hotspots. Its aggressive roadmap calls for suitable wireless chips in late 2006. These will support the 802.16e specification, as yet unratified as a standard, which will enable client mobility.
The trouble is, 802.16e includes many techniques taken from the mobile phone world that mean, as it stands, an 802.16e-equipped notebook cannot talk to an 802.16-2004 base-station.
Suddenly, the whole WiMAX enterprise begins to look a little shaky, in particular its role as a universal wireless system.
Been here before with Wi-Fi
Maloney likened the situation to the incompatibilities between Wi-Fi variants 802.11g and 802.11a. Both offer wireless connectivity at the same speed, but they are not directly compatible because they use different parts of the spectrum. That's not a problem, he told The Register, because Wi-Fi cards support both.
That's true today, but in the early days it wasn't. That's why 802.11b became the real standard in the public hotspot world, where even 802.11g has yet to establish itself. 801.11b and 802.11g operate in the same frequency, but a hotspot owner can be almost certain his users can access using b. He can't make the same assumption for g, even though almost all wireless notebooks these days ship with 802.11b/g and even 802.11a/b/g radios.
The implication for WiMAX is that the technology will be established with 802.16-2004, which will then be unable to cope with a growing number of 802.16e-equipped notebooks come 2006 (or so - this depends the accuracy of Intel's predictions).
Intel's answer is to "solve the problem in silicon". Miniaturisation will enable a single chipset to support both forms of WiMAX, allowing base-stations to establish connections to customer premises equipment and to roaming notebooks. There's a risk, however, that 802.16-2004 will become the baseline, as 802.11b did in the hotspot world, before 802.16e is put in place. Alternatively, the WiMAX market may take long enough to establish itself that 802.16d/e silicon is widespread by the time providers start rolling out the technology in earnest. But the way Maloney, for one, talks about how so many WiMAX trials are underway, roll-outs may come sooner rather than later.
3G to the rescue?
There's another possibility. According to Maloney, moves to combine WiMAX with the 3G-based high-speed data technology HSDPA (High Speed Data Packet Access) are already underway. He thinks thagt aligning the two standards is "technically do-able", and it could see 802.16e becoming a successor to 3G, at least from a data perspective, with 802.16-2004 emerging as the wireless successor to DSL.
That's fine, until someone tries to connect an 802.16e machine to an 802.16-2004 network and finds they can't, even though they're both promoted as WiMAX-compatible.
The logical solution would be clearly to separate WiMAX Broadband from WiMAX Mobile, but the use of the name implies a level of compatibility that isn't there, and too many companies will want to have their wireless cake and to eat it for the WiMAX name to be dropped from either incarnation.
WiMAX has been promoted by its proponents as a universal high-speed wireless data system, but incompatibility, however necessary, challenges this claim, and creates the kind of uncertainty among potential users and customers that can hinder its uptake. Combo products of the kind Intel's Maloney talks about may indeed ensure WiMAX's universality, but the standard faces fragmentation if bringing the two modes together is mishandled or simply not offered by every vendor. ®
WiMAX summit: 'Standards-plus' could harm 802.16 roadmap
WiMAX hype peaks
Tokyo to enjoy city-wide WiMAX
Qualcomm: WiMax isn't magic
T-Mobile launches WiMax net access for UK trains
WiMAX turns the screw on 3G
Motorola damns WiMAX with faint praise