Network World — Back in 2007, Nemertes Research predicted a "broadband access" gap that would begin affecting Internet connectivity by the year 2012.

Specifically, we noted that although core routing, switching and transmission capacity in the Internet were growing exponentially, access capacity (including both wired and wireless) was growing merely linearly -- meaning that the edges weren't keeping up with the core. More importantly, Internet traffic was also growing exponentially -- meaning that bottlenecks would soon begin to occur at the edges.

Turns out our predictions were spot-on -- just a little too conservative.

Fast-forward to 2010, and AT&T's recent announcement that it's plowing $2 billion into network capacity, primarily to handle wireless applications running on iPhones and the newly released iPad. Meantime, Verizon (VZ) has reportedly sped up its planned rollout of Long Term Evolution (LTE) to try and stay abreast of capacity issues.

And things are going to get a whole lot worse before they get better. Several trends are converging that will bring the wireless Internet capacity gap front and center to the attention of network managers. These include:

* Device convergence. The distinction between a wireless PC and a smartphone continues to blur, with devices increasingly serving both as computers and communicators. This, in turn, shifts traffic that might have previously traversed a wireline network onto the wireless network.

* Cloud computing. Increasingly, companies are looking at putting some or all of their computing resources in the cloud, increasing Internet traffic substantially. Combine this trend with device convergence (above) and the stage is set for an increase of cloud-based applications across wireless networks.

* Desktop virtualization. Although I'm not predicting as sharp an increase in desktop virtualization as server virtualization (for one thing, the cost savings aren't nearly as clear), companies are increasingly looking into desktop virtualization as a way to maintain control over corporate data. And once again, device convergence means desktop virtualization is increasingly driving traffic across wireless networks.

* Videoconferencing. Last year, we predicted (and saw) a dramatic uptick in the use of videoconferencing, both high-end telepresence suites and low-end desktop video. As companies adjust to a virtual workforce this trend will continue -- with users seeking to videoconference from wireless devices.

* Virtual workers. Applications such as social networking, unified communications and other collaborative technologies can unify far-flung workforces. As these folks increasingly move to wireless devices, these applications will increase demand on wireless networks.

The bottom line? These scenarios and applications will increase demand for Internet capacity -- and wireless Internet capacity in particular. That means network managers will need to pay close attention to the current and future data capacity of  wireless service providers. It will be a key differentiator in years to come.