For 3G Networks, the Forecast Calls for Rain:

New mobile handheld devices like Apple’s iPhone 3G and the newly released G1 from Google, which are designed to maximize multimedia applications of 3G networks, represent a great opportunity for mobile operators in the long term. More immediately, however, these new phones are creating migraine headaches, driving mobile network operators headlong toward a backhaul bandwidth crunch as 3G services generate much more traffic than their 2G and TDM-based analog mobile network predecessors.

From a consumer perspective, mobile phone users continue to expect exciting new features such as social networking on their handheld devices, but also want to see a decline in their subscription rates. But the current forecast calls for rain, since average revenues per user, or ARPU, the telecom industry’s favorite acronym, are actually on the decline precisely at a time when carriers need to invest in additional backhaul capacity that is capable of carrying the 3G services they’re banking on to generate new revenue streams. Before operators can book these new revenues, they must deploy the infrastructure at a time when ARPU is shrinking, which is making investment difficult. If operators cannot fund infrastructure investment out of new revenues, the only alternative is to borrow, which as we have come to learn, is now much harder, not to mention much more expensive even if credit actually can be obtained.

When looking at the actual numbers, it becomes frightening, particularly if bandwidth projections based on current bandwidth consumption by 3G networks users are correct. At present, many of the 400,000 cell towers in the U.S., for example, are served by only one or two T1s, delivering backhaul bandwidth up to only 3mbps. Yet within a year, 90 percent of backhaul networks will have to carry 30mbps, with the remaining 10 percent of the networks carrying as much as 100mbps. Again, this is based on current 3G phone projections combined with discussions I’ve had personally with the operators themselves on the subject of HSPA challenges.

These figures represent an increase of nearly 30-fold on current backhaul capacity, and are based on realistic usage scenarios. A mobile voice call uses only 7.5kbps, while a typical 3G data session runs at a persistent 3mbps, but only if the bandwidth is available.

Furthermore, with growing levels of mobile broadband Internet access, e-mail usage, texting and instant messaging (IM), peak levels of bandwidth consumption are being sustained longer. Even worse, growing use of video and mobile TV services is leading to more continuous consumption of bandwidth and at higher rates. Emerging high-definition mobile TV services stream at 250kbps, and at this rate, a cell tower with two T1s could accommodate only 12 users watching simultaneously at any one time, leaving no capacity left to make a voice call or access the Internet.

The natural question then becomes, how will mobile operators meet this rapidly proliferating demand for bandwidth without breaking the bank? One option, of course, is to deploy fiber, but unless fiber is already passing close to both the cell towers and the switching center at either end of the backhaul link, this is prohibitively expensive, and in most cases, can be ruled out. Alternatively, a poor choice is to continue on the current path and boost backhaul bandwidth merely by recruiting more T1/E1s. Given the huge number of T1/E1 circuits that would need to be turned on to meet predicted backhaul bandwidth requirements, this could work out to be even more expensive than laying fiber, even assuming there are enough copper pairs in the ground.

But before mobile operators become stuck between a rock and a hard place, they should consider a far superior third option — using their existing copper infrastructure just as many are doing now for their T1/E1s. Rather than paying for more T1/E1s, service providers can deploy carrier-class Ethernet in the First Mile (EFM) over copper, achieving immediate cost savings while boosting bandwidth per pair by a factor of approximately seven. While T1s run at 768kbps and E1s run at 1024kbps per pair, EFM over copper can run at more than 10mbps per pair. This means that backhaul commitments can be met without having to recruit anything near so many copper pairs.

In this perfect economic storm, you had better invest wisely. Your copper, after all, is your network’s BFF.