InfoWorld's Ephraim Schwartz argues that paid hot spots will be overwhelmed by cell data except in airports: I disagree with many of the premises on which Ephraim makes his argument, but I also don't think that I can prove myself correct until a few years have passed. Perhaps we need a bookie to get involved and set odds?
I'm of two minds on for-fee hot spots: the only way for them to succeed in the long-run is to be part of larger networks that charge fixed monthly fees and to offer significantly more bandwidth than cell data can possibly provide. Cell data is a great idea, but it's a shared pool of fixed spectrum with high costs attached. The carriers are all getting involved in Wi-Fi for reasons that contradict Ephraim's: they want to leach off data onto cheaper-to-operate hot spots to keep their voice networks able to handle more and more call volume.
A couple of fundamental disagreements:
Seamless switching a vague promise? It's a near-term reality, probably first surfacing in limited form in a few months, and then widely available by the end of 2004. I don't think it's very difficult to accomplish, as I've seen a demonstration -- that required client/server software -- at NetMotion Wireless's office and in the field with their equipment.
DSL runs at 600 Kbps to 1 Mbps: I've had 1.5 Mbps ADSL that did provide nearly 1.5 Mbps of download speed. There are folks in Snohomish, a tiny town with a small central office a few dozen miles north of Seattle, who have 8 Mbps downstream and 1 Mbps upstream ADSL. DSL can operate that fast if there's a market force pushing the telcos to upgrade their central offices to support better speeds. SBC is such a company and this might be part of the motivation they need. Technically, there's no reason for DSL to run as slowly as it does.
Finally, I'll just reiterate my point: ubiquity is trumped by speed and price. It's likely that an unlimited national roaming Wi-Fi plan will cost $20 to $30 added on to a cell subscription within a year. (You can already get all T-Mobile locations for $20/month and all Boingo Wireless locations for $21.95/month.)
The 2.5G and 3G rollouts aren't priced in the same class, and cell carriers have made it clear in recent weeks that you aren't going to see 3G speeds nationally til probably 2005 or 2006 -- if they decide to commit to it, which isn't a given yet. Remember, too, that no one is quite sure whether 3G speeds will be possible in interior spaces like hotels and buildings, just as cellular voice is now problematic.
My paradox still stands: you'll pay $20 per month for for-fee Wi-Fi service all over the U.S. as a flat rate, but you'll also find lots of free service. Cell data will be much more expensive, offer lower speeds especially in congested urban areas, and have limited coverage areas for the next couple of years or more.