Can MBWA compete with 3G?

Last December, the IEEE standards body created a committee to develop a Mobile Broadband Wireless Access (MBWA) standard. The idea is that this will fill the perceived gap between wireless LANs and 3G phone services, supporting high-speed IP data services to mobile users.

The new specification will be called 802.20 and should provide symmetrical wireless bandwidth rates from 1Mbit/s up to 4Mbit/s, if the spectrum is available. It will also work with vehicles travelling at speeds up to 250km/h. The technology is cell-based and should support seamless handover between cells.

Wait a minute, though - isn't this what 3G packet data services will provide anyway? Yes, it is, but the 802.20 group cites low latency - 10ms compared with about 500ms for 3G - as a differentiating feature, and one that may mean the kids can use it to play Quake from the back seat of the family car.

Another significant difference is that an 802.20 cell could cover an area within a radius of approximately 15km. This is certainly more than 3G can provide at the same data rate. But as with 3G, the effect of "cell breathing" means that as the usage increases, the cell coverage will have to be reduced, meaning the advantage will diminish.

The business model for 802.20 services may prove attractive because a base station dedicated to IP data services should be simpler than a standard 3G base station.

What about the available radio spectrum to carry the data signals, though? 802.20 will not use licence-exempt bands, but will be fitted into existing mobile allocations. The minimum bandwidth required for 802.20 is around 1MHz, because all the base stations will use the same frequency. Therefore, it appears that the only potential customers will be the existing mobile operators, which already own the rights to suitable spectrum. Will they be happy to set aside 1MHz for a data-only service?

There are other obstacles to be overcome in designing broadband mobile services because of the problems of multipath interference. There are two basic approaches: use a multiplexing system such as Orthogonal Frequency Division Multiplexing (OFDM), which is relatively insensitive to these effects, or use a scheme that can compensate for them. 3G uses the latter approach, but the indications are that 802.20 will use the former.

I have my doubts that the 802.20 group has the expertise to solve all these problems and come up with a technology able to deliver services that customers actually want. However, many new representatives from mobile operators and mobile manufacturers are now attending the group's meetings. There were 240 at the last gathering and the signs are that the number will continue to increase. Perhaps they are there to keep an eye on what is happening, rather than to help? In any case, large numbers of participants do not make for efficient working.

The 802.20 group says it will work with other groups like 3GPP and 3GPP2, to ensure the standard meets their requirements. It will also have to work with the IETF and, perhaps, the ITU. All this interaction is likely to mean that the target of delivering the specification within two years cannot be achieved.

In two years' time, broadband IP services will be common on 3G networks. If 3G services provide what users want, there might not be any market left for 802.20.

Have your say: reply to IT Week