Dropped calls and no service are a fact of life in cellular networks. They're usually the result of problems with the propagation of radio waves, which is how the waves get from Point A to Point B and back again. This is a frightfully statistical process if there ever was one. Lots can happen between those two points, and if the signal is damaged or otherwise fades to such a low level that the receiver can't detect and use (demodulate) it, that's that. No signal, no service.

Sometimes, though, dropped calls are the result of simply not having enough capacity on the cellular network, and this is usually due to having too few cells. The farther a signal goes, the less overall capacity the cell has, because there will be more users sharing the spectrum on that cell, and more range usually means less reliability regardless. We've got coverage, but not capacity. Why not just put the cells closer together, you ask? Good question. The answer is that cells are expensive, not only in terms of equipment, but also real estate costs, regulatory fees and lots of legal bills from local residents who love their cell phones but don't want to live near a cell tower. The laws of physics are at work here, too: The higher the signal is off the ground, the better it will propagate.

Wi-Fi doesn't have these problems. Wireless LANs have gobs of spectrum to use - as much as 689 MHz worth in the US, plus another 50 MHz for municipalities - and they operate over relatively short distances. This means we've got capacity, but not coverage. Sure, we can deploy metro-scale Wi-Fi meshes, but we'd still have lots of little gaps in coverage to deal with and highly variable performance due to motion and the number of other active users nearby.

So, cellular has coverage via big, expensive cells on licensed frequencies, and Wi-Fi has capacity on lots of little, inexpensive cells on unlicensed (free) spectrum. This, of course, begs the question: Why not merge the two into a single system and get the best of both worlds? Why not, indeed?

The interim step
Not so fast. There's an interim step we need to do first, and that's what's becoming known as fixed/mobile convergence (FMC). In its simplest form, FMC involves connecting the enterprise private branch exchange (PBX) to the cellular network, enabling cell phones to become extensions on the office telephone network. I expect this will ultimately happen via services offered by cellular operators, which will eventually realise that they can sell vast numbers of phones to the enterprise instead of one here and one there to consumers. Cellular operators also will be able to largely displace wireline carriers in the process.

Of course, the wireline guys will fight back by becoming mobile virtual network operators (MVNOs), cellular providers that lease their capacity from the cellular network operators, and offer their own FMC systems. And the enterprise can even go the do-it-yourself route and implement FMC via customer premises equipment, which is available from a number of vendors, with more on the way. We expect the MVNO strategy to become obvious in about a year, and the cellular operators to begin to move in earnest in about 18 months. They're in no hurry; they don't need to be, and this upgrade represents a major change to their core networks and services, so caution is in fact appropriate.

Key to all of these scenarios are dual-mode cellular/Wi-Fi handsets, which allow both voice and data access via both cellular and Wi-Fi services. Imagine using your cell phone as a cordless phone in the office, with unified dialing directories, calling features, a single voice-mail system (my No. 1 wish) and all of the resulting convenience. I've yet to meet an end user uninterested in this possibility.

Now imagine that the handoff of voice and data connections between cellular and Wi-Fi, at both the in-building enterprise and metropolitan scales, is transparent. I've been referring to this possibility as mobile/mobile convergence (MMC), since no connection to a PBX is required or necessarily implied. Indeed, I expect the PBX to gradually disappear as its services are replaced by the IP PBX, or, more likely, services provided in the carrier's cloud and administered by the one person left in the telecommunications department via a Web interface. Regardless, with MMC, we'll be able to unify the two wireless worlds I mentioned above and gain access to both coverage and capacity in a single handset supported by what appears to be a single service - you'll get one bill, anyway. Note that one radio just can't do it all.

All of this capacity, coverage and capability will eventually lead to the final step - fixed/mobile replacement (FMR). OK, I just made up that term. But think about it - if wireless can provide voice, broadband data, maybe a little video (granted, not HDTV), and can do it with the throughput and reliability of wired broadband, why won't we replace wire, just as so many have by making cell phones their only phones today? Why not, indeed?

Craig Mathias is a principal with Farpoint Consulting. This article appeared in Computerworld.