Fast and Spurious

When it comes to Internet delivery, the U.S. ranks either No. 2, 4, 7, 17, 34 or 44, compared with other countries, depending on who’s counting — and how.

After years of reports that America is falling further behind in the global broadband race, the divergent and contradictory rankings add to the perplexity of determining where the U.S. stands on high-speed Internet access — and if that standing even matters to domestic broadband users.

The long-simmering consternation about the stature of America’s broadband service bubbled up again last week in an upbeat report by the Information Technology and Innovation Foundation, a nonpartisan think tank that promotes tech innovation.

The study found that the U.S. has the second-lowest pricing for entry-level broadband access (after Israel); the world’s third-highest rate of wired competition (after the relatively small nations of Belgium and the Netherlands); and the seventh-best average access speeds on Earth (the average peak U.S. speed of 29.6 Megabits per second is significantly lower than South Korea or Singapore, but in the same range as Switzerland, Denmark and the Netherlands).

And at more than 80%, the share of U.S. homes passed by service at 100 Megabits per second or faster is among the world’s largest.

Last Friday (Feb. 15), the Federal Communications Commission issued its third annual Measuring Broadband America report, confirming that U.S. ISPs are delivering the access speeds they promote. Download speeds during peak hours average 97% of advertised speeds (cable 99%, DSL 85%), according to the study, which agreed with findings in the ITIF review (see page 3).

The ITIF’s report, The Whole Picture: Where America’s Broadband Networks Really Stand, is highly critical of the various measurement organizations, including the Organization for Economic Cooperation and Development (OECD), Pando Networks and “self-appointed watchdogs” that spread “indiscriminate criticism … that ignores successes that should be followed.”

The report also emphasizes that the U.S. “leads the world in the adoption of 4G/LTE [Long-Term Evolution] mobile broadband,” focusing on these fast-evolving wireless technologies as “credible” competitors to low-end wired broadband services and “a gateway technology for bringing broadband non-adopters online.”

Beyond the sheaves of data and the contradictory array of broadband rankings are more fundamental questions about the role of high-speed Internet access.

DOES SPEED REALLY MATTER?

How should broadband deployment be measured? What’s the value of comparing services in countries with vastly diverse infrastructures, government participation and patterns of media consumption? And does it really matter if the U.S. ranks No. 34 in download speeds or No. 44 in upload speeds over a four-year period (according to Ookla, a software diagnostics firm that relies substantially on self-reported data)?

The issue of price comparisons, including measures such as “household value” based on cost per Mbps, is fraught with potential inaccuracies. In the Ookla survey, the U.S. ranks 18th at $4.36 per month per Mbps, compared with Bulgaria (the equivalent of 54 cents per Mbps), Romania (64 cents), Lithuania (94 cents), Taiwan ($1.57) and the Czech Republic ($1.98). Yet the same survey is dotted with anomalies, such as the prices in Grand Forks, N.D. ($1.58); St. Peters, Mo. ($1.78); and Medford, Ore. ($2.19).

While these small-town U.S. services (some of them supported by government funding) barely skew the aggregated in larger markets, the levels add to the confusion as to what broadband service is really worth.

The vast variations also lead directly to two of the key factors cited by ITIF and other analysts: Speed and penetration. While the U.S. average maximum speed of 29.6 Mbps includes commercial services to businesses, schools and institutions, as well as residential access, “speeds in America are rising faster than elsewhere in the top 10 countries,” Richard Bennett, ITIF’s senior research fellow and an author of the ITIF study, said. More attention should be paid to “measuring the diffusion of high-speed deployments” rather than the installed base, he said.

In addition, the focus on speed obscures the reality of how much is enough.

“Are higher speeds worth the cost?” economist Scott Wallsten asked rhetorically at the ITIF seminar where the study was unveiled last week. “So far, there is no evidence that it is worth investing to achieve higher speeds.”

To support his view, Wallsten, vice president for research and senior fellow at the Technology Policy Institute, pointed out that over-the-top video service Netflix can effectively stream HD video at 1080p over a 5-Mbps network. The highest-capacity networks may be useful for intense usage, he acknowledged, but current levels support most foreseeable usage.

Nonetheless, the speed issue lingers, looking beyond streaming video to other applications where high speeds are vital. In particular, video games need speed to avoid data latency, assuring that massive multiplayer games give everyone the same real-time shot. Other applications, such as telemedicine — a long-promised benefit of high-speed networks require fast connections for high-quality diagnostics.

In his remarks, Wallsten also pointed out that not all customers want to pay “a lot more for higher speeds, which is often left out of the debate.” Nor do all customers value increased speed in the same way. Most subscribers can’t even recognize a speed increase in the broadband services they use, he said.

While the jump from 1 or 5 Mbps to 10 Mbps is certainly noticeable, customers are less likely to identify the difference between service at 25 or 50 Mbps. If that’s the case, why build even faster networks that will largely go unnoticed by most customers?

Network speed is often trumped by website and server capacity, especially as cloud delivery plays a bigger role. A usage surge at the host computer can frustrate even the fastest networks.

That raises even more questions about the value of the vaunted gigabit fiber networks that Google is deploying in Kansas City, Kan., as well as similar networks under development in cities from Chattanooga, Tenn., to Portland, Ore. Although the services are still works in progress, some analysts — including participants in the ITIF seminar — have suggested actual throughput may be in the 250-Mbps range.

Meanwhile, the policy debate over higher speeds persists. In the past month, FCC chairman Julius Genachowski escalated his pro-high- speed rhetoric, imploring Internet providers to build gigabit networks. In a blog on the Forbes website, Genachowski said gigabit networks “will certainly lead to unexpected new inventions” and allow more businesses “to take advantage of Big Data.”

Genachowski criticized National Cable & Telecommunications Association president and CEO Michael Powell for characterizing gigabit speeds as an “irrelevant exercise in bragging rights.”

The need for speed — and its value — will continue to be emphasized, as will the battle for international bragging rights. At last December’s B&C/Multichannel News OnScreen Summit, Time Warner Cable chief operating officer Rob Marcus said his company — the main rival to the Google Fiber buildout in Kansas City — “will provide” 1-Gbps service “if there is a demand for it.” (“ Marcus: We’ll Throttle Up Broadband in K.C.,” Dec. 10, 2012, page 4.)

“It will be interesting to find out whether there are applications that will take advantage of a 1-Gbps service,” Marcus said at the event.

The current focus on speed often supplants the long-held fixation on penetration into homes around the world. For nearly a decade, there has been increased hand-wringing over broadband penetration, even as cable operators and telcos spent billions of dollars on buildouts.

By various measures, the U.S. ranking in high-speed Internet penetration has dropped from the low teens to the low 20s.

Even as the U.S. broadband customer base expanded, the percentage of homes using high-speed Internet service has increased even faster in much smaller countries.

And as analysts have noted for years, the U.S., as a largely suburban country with customers spread over larger territories that are more costly to serve, must be assessed differently than other nations. One ITIF panelist said that while living in Singapore, his commute was mainly “vertical” — from the 18th story of his apartment building to a 21st-floor office a few blocks away.

Moreover, many of the most-penetrated nations offer a government subsidy encouraging broadband adoption. In some countries, local practices put building owners, not carriers, in control of the last hundred meters.

QUESTIONABLE DATA

Due to all of those factors, plus differences in pricing and service, conditions just aren’t comparable around the world. That’s another justification of ITIF’s critique of the arbitrary measurements of broadband success.

For example, the U.S. generates about 60% more network traffic per capita than Western Europe, according to a breakdown by independent analyst Bret Swanson of a 2009 report from Harvard’s Berkman Center.

As broadband deployment continues to accelerate — fueled by the addition of high-speed wireless services — the race to the checkered flag between nations will continue.

The FCC has started a move at OECD for “a more harmonized set of data … on broadband,” FCC International Bureau chief Mindel de la Torre said. Her suggestion recognizes the growing importance of a set of truly comparable metrics that reflect requirements and usage patterns in various countries.

But it also opens the door for even greater challenges, especially on issues such as cybersecurity. Future research must also consider growing use of wired and wireless Internet services by the middle class in fast-developing nations such as China and India.

Such challenges to the ultra-high-speed mantra — which isn’t part of the broadband roadmap in most other countries — add to the policy feud over the value of fast networks.

Paul Budde, an Australian telecom researcher, who for more than 20 years has conducted broadband studies for the United Nations and its ITU operation, among other clients, takes the long view about the value of comparative data.

“If you selectively use statistics, you can state whatever your agenda is,” Budde said. “It has more to do with the agenda that people want to push than with the real data.”

SPEED RACER

Traditional speed leaders: North Korea, Japan, Hong Kong, Netherlands and Denmark.

Average peak network rate in U.S.: 29.6 Mbps.

Average peak rate for other Top 10 nations: 37.9 Mbps.

Speeds improving faster in U.S. than in leading countries.

U.S. is installing more fiber-optic cable than all of Europe.

SOURCE: Akamai Technologies Q3 Report

HOW FAST ARE YOU?

Want to know how your broadband speed compares to others? Ookla, which continuously monitors broadband access, has a dynamic map (http://netindex.com) showing how fast data is moving in your vicinity. Click on the Speed Test (http://www.speedtest.net) to see how your computer or device is cruising along right now. You also might want to check the speed gauge on the FCC’s Broadband page (http://www.broadband.gov), which happens to be powered by Ookla. Actual speeds may vary.