Mobile broadband and phones....

I've spent some time recently thinking about the various flavours of wireless/mobile broadband and talking to assorted hopeful vendors.... Qualcomm/Flarion's Flash-OFDM, IPWireless' UMTS-TDD , Intel's favoured WiMAX , and assorted others, as well as various companies attempting to suggest that evolved "normal" cellular technologies like HSDPA could become viable ADSL/cable-rivalling broadband replacement technologies. The proponents of these technologies are already busy fighting among themselves . I'm wondering where the overall sector is going, and how long it'll take to get there.

There seem to be 3 main target markets

- laptops (data cards or embedded)
- fixed-wireless as an alternative to ADSL or cable broadband
- PDAs and phones

Laptops

The clearest demand (and perhaps the smallest ultimate addressable market) is in laptops. Following on from WiFi's success (especially post-Centrino), and reasonable uptake for WCDMA and EV-DO data cards, it is something of a no-brainer to suggest customers want faster, cheaper, more reliable and ubiquitous connections for their notebooks. There also isn't that much of a large market for people wanting to run VoIP from laptops, so the thorny issue of cannibalisation of voice revenues doesn't really matter here.

However, there are still plenty of problems around indoor coverage, which is generally lousy for all these technologies. I'm also very skeptical of the idea of embedded EV-DO / HSDPA capability being shipped by the manufacturer - who wants to tie themselves to a specific service provider for the 3-year life of a laptop? I shudder to think of the software hassles if you want to change to another network....

Broadband replacement

Is broadband wireless a DSL replacement? Hmmm. We've been here before . A million-odd customers in the US and 5m worldwide since about 1997. Big deal. Sure, there are a couple of early deployments of wireless broadband that are great for sparsely-populated, low-competition areas in wealthy countries like Finland and New Zealand, where the local fixed/cable operators haven't been forced to roll out networks. And TDD and WiMAX seems to have found a promising niche as backhaul for on-train WiFi, while a few brave operators are using fixed WCDMA for home data users.

I'm entirely unconvinced this is a mass-market option, however.

The main problem? Copper (or fibre) is cheaper than spectrum. The other problem? It's usually easier to get planning permission to dig up a street, than to put a new base station on a roof. And most existing cell sites don't have enough backhaul capacity (or power) to turn them into the equivalent of a DSLAM or cable head-end. Most European 3G base stations have an E1 (2Mb/s) backhaul pipe. Adding capacity is complex, expensive and time-consuming. So while fixed broadband is already driving down the bandwidth curve... 512k, 1M, 2M, 8M, 24M, 45M, 100M, I've even heard plans for residential Gigabit Ethernet... the wireless options are still mostly at the stage of shared 2M capacity amongst an entire cell's user base.

Overall - it's another niche. At some places and some points in time, wireless (of any flavour) might temporarily compete with fixed pipes, but it's not a massmarket urban-area solution. "Everything" certainly won't be wireless, despite the hype. Ever.

Handsets

Lastly, phones. Here, the argument is more WiMAX/TDD/F-OFDM vs. 3.5G (EV-DO, HSDPA). Will "3rd-party wireless IP" free our phones from the tyranny of cellular operators' data services? Clearly, there are already phones on sale which include EV-DO and WCDMA, with HSDPA in the pipeline. There's WiFi in some phones already .

Will we all be using TDD- or WiMAX-phones in urban areas instead? Not this decade, no.

Apart from all the previous arguments about indoor performance, base stations and backhaul, there are a whole range of issues in actually creating phones with these technologies in. First off is the problem of convincing the major chipset suppliers (especially TI) to support them. Then there's a whole spectrum (pun intended) of RF problems in designing-in and testing new technologies that work at a wide range of frequencies. Then there's battery life - and even heat dissipation (remember how warm your modem gets? fancy a phone with fan?).

And the software layer - an IP-only phone will need a firewall and maybe anti-virus, which means an expensive "open" OS. And a whole bunch of complexity around the user interface and how it behaves. In general, adding ANY new software to a phone is much more involved (and time-consuming / costly) than the majority of observers expect. It's not just a case of saying "it only needs an extra $10 chip - what's the problem?"

Conclusion

This post isn't intended as an exhaustive round-up of wireless broadband technologies. But it highlights some of the practical problems, and is intended to deflate some of the hype. As usual, there's a couple of good stories (eg rural broadband, cellular network backhaul, road warriors' laptops) obscured by mountains of marketing guff and rampant over-optimistic speculation. I'm broadly optimistic on the laptop-broadband sector (I'd use it) but the "residential fixed broadband replacement" and "handset cellular replacement" sectors don't really fly in my view.