Net neutrality for wireless: The few versus the many

[1]FCC Chairman Julius Genachowski laid out [2] his priorities for the wireless industry at the CTIA Wireless IT & Entertainment event last week. He wants to:

• Bring more spectrum to market to handle rapidly increasing demand for wireless data
• Remove red tape to allow wireless carriers to expand networks faster
• Conduct the regulatory process at the FCC more openly and on a fact-based, data-driven basis
• Codify and enforce net neutrality with special considerations to wireless

I am sure the wireless industry welcomes the first three priorities of the new chairman. They represent a welcome and overdue recognition of the situation we are in--more than 270 million American wireless subscribers and more than 42 million of them using smartphones to access the Internet. While the discussion continues about the need for the codification of net neutrality for wireless, it is very encouraging that Chairman Genachowski has recognized that wireless networks deserve special consideration. Some examples:

1) Wireless data networks that are available to most Americans have only modest throughput.

Today's technology allows CDMA network operators (Verizon Wireless and Sprint Nextel) to provide a theoretical maximum throughput of 2.4 Mbps to 3.1 Mbps and UMTS carriers (AT&T Mobility and T-Mobile USA) a combined theoretical maximum of 3.6 Mbps. For simplicity, let's assume the throughput of a particular cell site sector to be 3.6 Mbps. This theoretical maximum is achievable if only one person uses the cell sector and is standing next to the antenna. The farther the person is from the tower or the more mitigating circumstances there are in between, ranging from interference from other sectors to mundane circumstances like leaves on trees, the less throughput available to an individual.

2) Wireless networks are a shared resource.

The throughput is shared among all people actively using the network at any point in time.

3) Adding capacity is not always possible due to limited spectrum availability.

Unlike wireline, where you can always put another fiber cable in the ground or light up another strand of fiber already laid, wireless service providers can't manufacture more spectrum.

For most applications this does not pose insurmountable problems because a subscriber uses the bandwidth for only a short period of time, the duration of a call or while loading a Web page. Such usage patterns are easily sustainable on current networks and are generally described as "bursty" traffic.

Complications start when we move from bursty traffic to streaming. With streaming, a constant amount of throughput is needed by the subscriber to maintain the connection and enjoy a satisfactory quality of service. The constant throughput cannot be used by anyone else. Streaming audio from an application like Pandora or Rhapsody typically uses about 40 Kbps and can support around 90 concurrent users (a theoretical best case scenario of 3.6 Mbps divided by 40 Kbps) in a cell sector. Wireless carriers allow and even sell streaming audio, which would indicate that there are significantly fewer than 90 concurrent listeners in a cell sector.

It gets difficult when the mobile subscriber uses streaming video in general and with an uncapped video frame rate in particular. Take the Slingbox, a popular device that lets people watch their home television service away from home on a laptop or mobile device. The quality of the picture depends on the uplink speed from their home Internet provider (between 384 Kbps and up to 10 Mbps) and on the download speed of their wireless carrier. In the best case scenario nine people (3.6 Mbps divided by 384 Kbps) can watch TV concurrently via their Slingbox, effectively preventing any other activity. Without the ability to manage the amount of throughput allocated to these nine individuals, other mobile users in the sector will be out of luck.

Currently, wireless network operators do not throttle data throughput, but this exposes them and consumers to a negatively impacted user experience caused by only a few heavy users. Traffic shaping could safe guard against the general slowdown of the data throughput in a cell sector for every consumer, but negatively impacting the viewing pleasure of nine subscribers to enable all users in the sector to receive service. This dynamic, real-time dance of RF engineering would be a forbidden under a purist net neutrality approach.

Another implication of a purist net neutrality rule that would negatively impact wireless subscribers' service is if one person uses their wireless modem to power their Web server--and yes it has happened.

How big is a cell sector? Anywhere from the size of the city block that houses Madison Square Garden in Manhattan, to the size of Capitol Hill in Washington, D.C., to several square miles of corn fields in rural Nebraska.

Considering the relatively fragile nature of RF networks and the reasonable expectation of a majority of subscribers that they can make and receive calls and download the applications they are paying for, it is only a matter of fairness to the majority of wireless consumers that carriers be permitted to allocate network throughput in a fair and equitable manner among all customers.

Roger Entner is senior vice president and head of research and insights for the Telecom Practice of Nielsen. For ongoing insights from Nielsen on this and other topics, visit NielsenWire.com [3].