One thing we love to do in high tech is invent new terms, especially acronyms, but new terms of any form really appeal to us. I must confess I find the terminology of technology to be one of its paramount appeals. Of course, new terms sometimes fall into pervasive use without adequate definition, leading to confusion, but making the originator or purveyor of such terms feel great nonetheless. Sometimes terms fall into disfavor almost as rapidly as they appear.

Two terms floating around right now are access and distribution. Access is most often used for "Internet access," implying the connection between a user or point of presence and the rest of the Internet, usually represented by and often called a "cloud." Distribution, on the other hand, is the mechanism that provides the connection between a given user and said point of presence. All of this is very interesting in the context of wireless because it defines what is certain to be one of the key battles in the evolution of wireless communications.

To illustrate this, consider two different models for wirelessly connecting to the Internet. The first, based purely on access, is to use a relatively long-range connection between the client and the entry point to the Internet cloud, akin to the way cellular works today. This is how most people envision 3G or WiMAX, for example, being used to access the Internet. The problem with this model is that it uses fairly large cells, and thus doesn't have good frequency-reuse characteristics for a large number of users and/or high bandwidth connections - limiting the number of users and/or the speed of their connections, thus driving up prices. But it does allow a user equipped with a mobile access connection to carry this connection everywhere the appropriate infrastructure exists.

The alternative would be to bring the Internet via a wired or wireless point-to-point connection to a point of presence, and then use a distribution mechanism to provide the final link to the user. An example here might be to use WiMAX or some functionally equivalent terrestrial microwave link to connect to a point of presence, and then use a Wi-Fi access point (again, for example) to provide the infrastructure end of a distribution network. The other end would be, of course, notebooks, PDAs, or handsets equipped with Wi-Fi capability. This allows a fair number of users to share the otherwise expensive wide-area broadband connection, and also largely solves the problem of getting wide-area signals to penetrate a building by avoiding it altogether.

Which model is better? That depends. It would be great to carry only one radio and have it work everywhere. But this is something of a pipe dream, since it's unlikely we'll see a single wide-area technology deployed everywhere. We also have a problem with building radio penetration as noted above, but we might be able to address this issue via repeaters or in-building microcells, as is done with cellular today.

On the other hand, using Wi-Fi as a distribution mechanism also makes a lot of sense, but you might need a lot of access points to make this work. We can also deploy Wi-Fi meshes or lots of access points in a metropolitan area, in effect using Wi-Fi as an access technology. Wi-Fi is certain to be pervasive, and has the potential for very high throughput, so why not just settle on that?

The bottom line is that both models will continue to flourish. Keep in mind that multimode clients (and, eventually, multimode clients implemented with software-defined radio, or SDR) will mitigate this issue to some degree. One of the really exciting possibilities for future wireless communications systems is that we may see wireless services disappear quite literally into the ether, and not have to worry about having the right radio. It could eventually all be quite automatic and transparent. While our primary goal in wireless communications remains to narrow or even eliminate the behavioral and performance differences that exist between it and wire, an equally important goal is to improve ease of use so end users can do more communicating and less network-oriented thinking and planning. Oh, and there's another term - SDR. We'll cover that in a later issue. And maybe a few others as well.