I was a little hesitant to sit down to write this column because I'm going to be talking about something that you won't be able to buy for a couple of years yet. But 4G is getting a lot of buzz in the press and elsewhere, and this is a set of technologies and services that deserve to be on your radar screen. If nothing else, 4G will bring wide-area wireless further into the broadband age, with more LAN-like capabilities than have previously been available in wide-area wireless systems -- more throughput, to be sure, as well as an all-IP infrastructure that further narrows the gap between cellular and wireless LANs.
First of all, what is 4G? The International Telecommunications Union (ITU) defines it as downlink (from the cell to the subscriber) throughput of 100 Mbps or more, and corresponding uplink speeds of at least 50 Mbps. This makes sense from their perspective, as they defined 3G as downlink throughput of at least 144 Kbps and up to 2 Mbps. But note, first of all, that these are peak speeds, and your effective performance will be much lower.
This is due in part to what throughput carriers are really willing to provision, the quality of the radios and antennas on each end, the distance between transmitter and receiver, the nature of the radio channel between the two points (intervening objects -- buildings, etc.) and the number of users attempting to share the channel at any given moment. In short, your mileage will indeed vary, but you're not going to see throughput anything like 100 Mbps. I think 2-4 Mbps is much more reasonable; less, certainly, but still a lot more than the 500-700 Kbps typical in today's high-end 3G services.
Because of this misleading gap between specification and reality, I personally take issue with the ITU's definition. First of all, I don't think 100 Mbps is reasonable in wide-area services because the cost of the spectrum utilized for cellular is really, really high, the price set by the results of government-sponsored auctions. The latest cellular auctions in the U.S. just ended and will raise about $13.7 billion for the Treasury. The carriers need to recoup this money somehow, and that somehow will be through the fees they charge us.
But imagine what they would have to charge if they gave any of us real 100 Mbps service – the price could be astronomical, given the number of ordinary phone calls that could simultaneously fit into that same spectrum. Today's data prices of $80 or so per month are, I believe, the result of the carriers' matching capacity to demand. If they were to cut prices, they simply would not be able to service the resulting crush of customers without degrading their voice services. So it seems a better strategy for 4G is to offer lower speeds, but more capacity for more users as the bargain.
Thus, my definition of 4G is a little different from the ITU's. Farpoint Group's definition is an all-IP, wide-area (range of at least several kilometers), broadband (throughput of 2-4 Mbps), wireless, mobile (obviously) system with full support for time-bounded traffic. This means that 4G will move us into the era of wide-area VoIP services, finally leaving the TDM/SS7 world behind once and for all, and certainly tracking the thinking (and deployments) at work in the evolution of landline services today. And that, to me, is the real value of 4G – building a next-generation wide-area wireless service that is philosophically and technically compliant with the future of telecommunications as a whole.
Now the bad news -- 4G is more than five years away in terms of real deployments, and it may be much longer than that. Sprint's proposed mobile WiMAX service could indeed be 4G, but the other major carriers won't move in this direction for some time. They still have a long way to go in their 3G deployments and the recovery of the costs therein. In the interim, it may just be public-access Wi-Fi deployments that grab the headlines. Following my definition above, I think you'll agree with my very surprising conclusion here: Public-access Wi-Fi networks are in fact the first 4G systems.