Everybody Hertz: The Looming Spectrum Crisis v

Table of Contents

  1. Summary


Net neutrality may be the headline-making broadband battle in Washington these days, but there’s a far bigger battle brewing at the capital — conflict over access to spectrum.  Federal Communications Commission chairman Julius Genachowski admitted during a visit to GigaOM’s office last month that he was surprised by the amount of time he spent learning about and dealing with spectrum policies. But as network-connected devices such as data cards, smartphones and network-connected gadgets from refrigerators to cars proliferate; the demand for spectrum could outstrip the supply.

The electromagnetic spectrum is the invisible highway system that carries bits of information from your mobile device back to the Internet. But mobile devices aren’t the only vehicles on that highway; radio waves, digital television, your baby monitor and Wi-Fi also travel this highway system in different lanes, known as bands.

Unlike the wired web, where innovations in fiber optics are leading to ever-expanding capacity, the mobile highway is relatively fixed. Each band of spectrum contains only so many megahertz, divided up by regulators, leaving the highway only so wide. To add capacity, either regulators need to make more swaths of spectrum available for mobile broadband, in effect creating new highways, or technologists need to figure out how to deliver more bits per hertz, the equivalent of cramming more commuters in the highway lanes. Alternatively technologists can figure out how to use a wider range of the available spectrum, such as the 60 GHz band, which has less ideal properties for long-distance data transfer. But that’s a risky endeavor.

Most experts believe we will have to rely on both technologists and regulators to meet the demands that consumers are currently placing on the mobile broadband infrastructure operated by carriers, and we’ll have to do these things sooner rather than later.

Independent analyst Peter Rysavy estimates that even if the government allocates new spectrum for carriers, it wouldn’t be available for five to seven more years thanks to a slow-moving political process. At the current rate of mobile broadband usage growth we will run out of spectrum well before that time.

Case Study: AT&T’s Network Problem

As an example of the problem’s magnitude, consider AT&T. The carrier is a poster child for how increasing demand for mobile broadband can adversely affect a network. After it received the exclusive right to offer the iPhone in the U.S., the carrier saw its data traffic grow by 5,000 percent. In December, 1 percent of smartphone customers used 20 percent of AT&T’s network capacity, while the top 3 percent of smartphone users used 40 percent. In the last year alone, AT&T saw its broadband traffic double. During its fourth quarter earnings call, AT&T said it was deploying three or four HSPA carriers, which is equipment that uses about 10 MHz of spectrum. This means in some markets AT&T is already using almost 50 percent of its spectrum allocation in some markets  simply delivering 3G data. If its traffic doubles again, AT&T has a problem.

No wonder AT&T told the FCC that it should allocate up to 1 GHz of total spectrum for cellular traffic. AT&T is not the only carrier facing this shortage. The CTIA notes that its membership believes the government should allocate 800 MHz for mobile broadband, and Clearwire told the FCC that in the near future, when individuals will use applications that require access to and transfer of 10 GB, 15 GB, or even 20 GB of data, download speeds will be diminished to dial-up equivalencies unless more spectrum is made available. And as newer generation superphones and even smartbooks hit the network, the problem worsens.

How can the problem be solved?

In an earlier report, I talked about pricing plans and mobile offload as solutions to the capacity crunch that operators face. But neither one of these options alone will solve the problem. For example, data from Cisco suggests that Wi-Fi offload will divert only a quarter of mobile web traffic by 2014, leaving a whopping 3.6 exabytes of data to be delivered via mobile broadband each month. Clearly, shifting traffic off the network won’t be enough to ease the coming gridlock.

The solution isn’t just about adding more spectrum, either.  As GigaOM Pro partner analyst Chetan Sharma put it in a recent report:

It is apparent that to achieve next generation broadband speeds like 50-100 Mbps, new contiguous spectrum is needed. However, it will be a mistake if the dominant solution for the broadband capacity crisis is more spectrum, for the following reasons:

  1. There isn’t enough spectrum, especially the right spectrum
  2. It takes 7-10 years to procure the spectrum for wireless use
  3. By focusing on spectrum only, we will be just postponing the current crisis

So, what’s that innovation look like? One huge shift already taking place is carriers’ movement to new network technologies, such as the shift from 3G technologies (like UMTS and HSPA) to 4G technologies (like  Long Term Evolution, or LTE). In addition to lowering the cost of delivering bits, these technology shifts  also boost spectral efficiency, or the bits per hertz that an operator can force through its network (see Fig 2). In fact, the ITU recently made the ability to transfer more bits per hertz an integral aspect of meeting the standards for next-generation wireless technologies. This way operators boost speed and capacity using the spectrum they already have.

Comparison of Downlink Spectral Efficiency
Comparison of Downlink Spectral Efficiency (click to enlarge) Source: Rysavy Research
Comparison of Uplink Spectral Efficiency
Comparison of Uplink Spectral Efficiency (click to enlarge) Source: Rysavy Research

Another way of achieving spectral efficiencies is what’s known as MIMO, which stand as for multiple-input and multiple-output. Essentially, MIMO uses multiple antennas to boost a tower’s ability to broadcast a signal, and it can as much as double capacity, depending on the configuration. Vendors pitching LTE and HSPA gear are working MIMO into their equipment. Ken Biba, an analyst with Novarum, a wireless research firm, says MIMO works best if it’s applied both at the base station level and inside the handsets. However, because that would put more silicon inside handsets, it would make them more expensive for consumers.

Another technology that might help is software-defined radios, such as those Alcatel-Lucent launched in the first week of February. These radios allow a carrier to allocate certain network technologies, such as UMTS or HSPA, in different bands of spectrum through a remote software upgrade. That means it’s not a real-time shift, but it can be done in quickly enough to respond to anticipated crowds or developing congestion points, such as a high school where iPhones have become the hot new toy. A carrier can allocate all of its spectrum bands to serve 3G data if it needs to, or split the spectrum to allocate 5MHz for 2G and the rest for 3G. Such flexibility allows the operator some granularity so it can match its spectrum assets to the demand.

Adding more spectrum

While carriers are exploring many of these options, they’re also begging for more spectrum resources from the U.S. government. The government is likely to oblige, too, although the details are still up in the air over which blocks the FCC will allocate and how long that process will take. But many analysts and industry folk are unified in suggesting ways that the government can implement spectrum allocations to make the most of the airwaves.

The first suggestion is to unify U.S. spectrum blocks for various technologies with those in other countries. For example, U.S. carriers can’t take advantage of Alcatel-Lucent’s new software-defined radio because the spectrum bands in America aren’t the same as those used in Europe for UMTS and HSPA.

Click to Enlarge. Source: Rysavy Research
LTE Spectral Efficiency as a Function of Radio Channel Size (click to enlarge) Source: Rysavy Research

Another suggestion is that the government offer spectrum in 20 MHz blocks rather than the smaller bands licensed today. The advantages are that larger blocks transmit data more efficiently (see chart below), and it’s easier to build hardware that tunes into larger bands of spectrum. The challenge is finding large enough chunks that can be reallocated for mobile broadband, and then gaining the Congressional support to force the current occupant of that spectrum to give it up. That won’t be an easy task.

Network neutrality regulations are the final area in which the government’s actions will impact spectrum, and while they are not about increasing the overall efficiency, they will likely affect how carriers manage the looming specter of network congestion. Currently, the FCC is trying to create formal regulations that will mandate that wireless network operators must follow the same rules that wired broadband ISPs will follow; namely the FCC wants to prevent wireless carriers from discriminating against any legitimate traffic on their networks. The rules would let a carrier interfere with traffic under the guise of reasonable network management, but it’s still unclear as to how that will be defined. That lack of clarity (the rules won’t be set until Spring and even then lawsuits will challenge them) means it’s hard to predict if net neutrality rules will contribute to the spectrum crisis or if they will have little impact.


There are plenty of alternatives before we declare an airwave Armageddon. The quest for spectral efficiency through technologies such as MIMO and moving from 3G to 4G will net big gains for operators. As engineers cram more bits into each hertz, policy makers like Genachowski and members of Congress evaluating our nation’s current spectrum use, will be weighing the cost of taking spectrum away from existing users, such as the broadcast television business or the Department of Defense, and reallocating it to mobile broadband. That policy discussion will involve bitter fights over the right to use the airwaves, whether or not the spectrum allocations should be auctioned off to the existing carriers, AND whether it should be unlicensed or licensed. While some argue that Congress shouldn’t get involved, policy issues will be an important piece of the solution as well.  Engineers are doing their part, and Congress will need to do so, as well.

  • By giving out spectrum too soon, industry won’t have the opportunity to learn to thrive within its means and let new technology and business innovation show the way to handle the increased data consumption
  • Access Report

    Available to GigaOm Research Subscribers

    Subscribe to
    GigaOm Research