The green mile

In LMDS and other high-frequency formats, we're talking about an up-and-coming access format that's billed as having fiber-like reliability and similar bandwidth potential.

An industry contact (one who prefers to remain cloaked) recently floated an intriguing theory past me about last-mile economics: He believes the cost curve for deploying fiber in the metropolitan network is plummeting so quickly that alternative technologies - broadband wireless in particular - likely will be used only as short-term fillers.

Immediately my wireless bias kicked in, and I dismissed this guy's hypothesis outright. In LMDS and other high-frequency formats, we're talking about an up-and-coming access format that's billed as having fiber-like reliability and similar bandwidth potential. Broadband wireless can, its proponents say, be deployed quickly and cheaply from rooftop to rooftop. No digging up the streets and no feeding fiber into conduit, so presumably fewer technicians are involved. Its only identifiable obstacles are rain fade (now solved, most say), line-of-sight limitations and roof rights.

My cloaked friend's view is that a network buildout plan should be based on the ultimate goal of taking fiber all the way to all the buildings a carrier wants to light and that cost models will eventually support that approach. Broadband wireless might be used in certain applications in this scenario but only as a high-tech patch until the carrier can afford to push fiber all the way. Once it can, the broadband wireless gear would be yanked out or simply become obsolete.

I brooded about this for a while and queried several analysts and service providers to see what they thought. I steered clear of the broadband wireless vendors themselves because I could fairly easily predict their answers. I decided that this argument does have a certain degree of merit but that it does not present the complete picture.

Yes, fiber is the ideal medium for connecting the network to the customer because it's a natural extension of what likely will be an already fiber-based network core. Yes, if the economic metrics for deploying last-mile fiber were the same as those for deploying last-mile wireless, fiber would be the immediate choice. Fiber is physical, wireless is virtual. Broadband wireless is a fast fix that can be put in place while the fiber is being extended.

But it's at that point I part ways with this theory. I believe two applications will prevent broadband wireless from ever becoming obsolete as an access technology.

First, if broadband wireless is used initially to put a building on-net and said building is eventually fed with fiber, the radio gear can continue to act as a spoke off the fiber network, even if it doesn't serve any customers in that particular building. That way, the broadband wireless antenna continues to function as a fiber extension.

Second, if there are no un-fibered buildings within range to take advantage of the first application, then the broadband wireless equipment can be taken down from its current location and redeployed in an area still lacking in fiber.

Perhaps my understanding of this technology is elementary, but it's based on hours of explanations from vendor and carrier engineers who are developing or deploying this equipment. Thus I refuted my contact's theory based on the assumption that the communications service providers that have invested millions of dollars in spectrum licenses and are about to make similar investments in equipment are not operating under short-term assumptions.