By Craig Mathias

There's little doubt that 4G is the path to the future of wide-area wireless communications, often lumped under the general heading of cellular, wireless, or mobile data services. Indeed, all major cellular carriers are now deploying 4G in some form, most importantly LTE. LTE has been carefully specified and quantified by industry bodies, and the marketing hype is now in full bloom.

But it may come as a surprise that LTE, universally marketed as 4G, may not be 4G at all. There is, in fact, an official definition of 4G from the International Telecommunications Union (ITU), often the arbiter of what is what in the world of wireless. The ITU, for example, quantified 3G as wireless throughput of up to two Mbps, and qualified six different technologies as meeting their specification. The ITU's definition of 4G, though, specifies throughput starting at 100 Mbps (!) and going up to a gigabit per second. Two technologies, LTE-Advanced and WirelessMAN-Advanced, have already qualified under this definition as 4G.

Needless to say, however, such levels of performance are unavailable anywhere today, and Farpoint Group believes it will be many years before such services become common. The key stumbling blocks here are the availability of the large blocks of free radio spectrum and the massive amount of money required. What such services might cost subscribers is also impossible to estimate today, and wireless carriers are, after all, still in the middle of their 3G deployments and won't be in a position to deploy "official" 4G for many years yet.

This reality, though, hasn't stopped those carriers from marketing LTE, WiMAX (a 3G technology, by the way, according to the ITU), and even HSPA and HSPA+ as 4G. Such is not entirely unfair, as performance for each of these can easily exceed the two Mbps of the ITU's 3G definition. Other unofficial designations, including 3.5G, 3.75G, and 3.9G have consequently been used to describe these technologies.

Even if these services aren't "real" 4G, they still have potential that is nothing short of mighty: HSPA+ technology has been specified all the way to 168 Mbps, and LTE to 300 Mbps (in both cases referring to downlink speeds). So calling these services 4G doesn't seem so outlandish, even if it's based on zealous marketing and not the blessing of a standards body. Moreover, effective downlink throughput of up to 10 Mbps has been reported, which is pretty good no matter what label one might like to apply.

Bandwidth Barriers

Before we get too excited here, however, it's important to point out that higher levels of throughput, which has been key element in the transition from one cellular generation to another, is by no means guaranteed. The intrinsic capabilities and limitations of a technology define a theoretical upper bound, but in practice, the range of variables gating real-world performance is enormous: the technical details of a specific implementation, a carrier's specific deployment, and network loading all impact realized application-layer performance.

Carriers might, for example, deploy a given technology less densely to save money, but such reduces capacity. The distance between the endpoints of a wireless connection is also a key variable in determining reliability and thus throughput, primarily due to natural signal fading which increases with distance. And it's a sure bet that performance numbers like the 10 Mbps noted above were obtained on lightly-loaded infrastructures. Farpoint Group believes that a key reason we have not seen an LTE-based iPhone, for example, is that the carriers would simply be overwhelmed by an onslaught of easily-disappointed customers. Such may still be the case once an LTE-based iPhone appears regardless; again, it's all about coverage and capacity, and critical mass. Enough availability and capacity for mission-critical users to depend upon can take quite some time to realize.

Next Steps For Enterprise Users

So, what strategy should IT and network managers pursue today? First and foremost, it is important that IT planners not let application requirements get ahead of reality. In other words, what we - or the carriers - call the technology doesn't really matter at all. What does matter are the capabilities of 4G (and 3G) that accrue directly to the user, including geographic availability, price, coverage, capacity, effective throughput, and, most importantly, the ability to meet application requirements, which must always be specified conservatively. And these elements can be evaluated independently of how a given service is marketed.

The bottom-line issue for end-users is ROI: does the subscribed service meet the needs of users and applications? Can payback for an investment in wireless services be seen over the useful life of a device and its contract (typically, two years)? And the bottom line for IT is that it really doesn't matter what the service is called -- if a carrier can cost-effectively meet the current and anticipated needs of its base of users and applications, then that's as good as it gets.

Perhaps, then, more time should be devoted to making sure the carrier is aware of end-user needs and making the necessary adjustments, and less time spent debating about how a wireless technology should be classified.

Return to the 4G Special Report Home Page