Wednesday, December 17, 2014

At least 59% of U.S. Households Can Buy 100 Mbps Internet Access

At least 59 percent of U.S, residents could buy Internet access service at a minimum speed of 100 Mbps, at the end of 2013, according to a report by the U.S. Department of Commerce. That probably is worth keeping in mind the next time you see or hear it said that U.S. high speed access is “behind,” or “worse” than some other countries.

But that statistic raises questions. Just how much bandwidth do most U.S. consumers really need? And how much bandwidth is required so that user experience is not impaired?

Without in any way suggesting there is anything wrong with a “more is better” approach, or suggesting that demand is growing, what is needed to support the everyday applications that users on fixed or mobile networks typically undertake?

Basically, peak bandwidth requirements are driven by the number of people sharing a single account, the amount of time that each user spends using the Internet, the number of simultaneous users and the types--and numbers--of apps they use when online. Application bandwidth requirements also are an issue.

The Federal Communications Commission currently suggests a single user can manage without difficulty, when watching entertainment video, at about 0.7 Mbps for “standard definition steaming videos.”

A single user requires 4 Mbps for high-definition quality streaming video.

Speeds between 1 Mbps and 2 Mbps are enough to support as many as three simultaneous users interacting with email, surfing the web or watching standard-definition streaming video.

But requirements are more complex as higher-bandwidth applications and the number of users grow. When four or more people are using a shared connection, and HD streaming, video conferencing or online gaming are used by multiple users, requirements can jump to 15 Mbps fairly quickly.

Netflix, for example, recommends speeds between 0.5 and 25 Mbps, depending on image quality, with 4K video imposing the highest loads.

As a baseline, the Australian National Broadband Network’s goal is to supply all homes and businesses with downstream speeds of at least 25 megabits per second, by 2020, with a majority of premises in the fixed line footprint getting downstream speeds of at least 50 megabits per second.

But NBN also will be using satellite delivery and fixed wireless platforms that will not deliver such speeds.

So demand assumptions matter. Single-person households likely will not be an issue, but multi-person households represent more complex challenges. On average, Australian Internet households represent about 2.1 people per household.

A new study by Ofcom, the U.K. communications regulator, found that “access speed” matters substantially at downstream speeds of 5 Mbps and lower. In other words, “speed matters” for user experience when overall access speed is low.

For downstream speeds of 5 Mbps to 10 Mbps, the downstream speed matters somewhat.

But at 10 Mbps or faster speeds, the actual downstream speed has negligible to no impact on
end user experience.

Since the average downstream speed in the United Kingdom now is about 23 Mbps, higher speeds--whatever the perceived marketing advantages--have scant impact on end user application experience.

Some 85 percent of U.K. fixed network Internet access customers have service at 10 Mbps or faster, Ofcom says.

So paradoxically, as there is increased focus on gigabit networks and services running at hundreds of megabits, more commonly, there remains little evidence, so far, that it actually makes a difference to most end users, in terms of better user experience.

What faster speeds might do is turn a spotlight on all the other elements of the delivery chain that impede experience, ranging from devices and Wi-Fi to far-end servers and backbone network interconnections.

None of that is obvious to end users, and little of the reality matters for marketing platforms, which suggest “more is better.” Of course, in the end, more is better. Up to a point.

No comments:

Generative AI Will NOT have the Impact Many Expect

Generative artificial intelligence, to say nothing of machine learning or neural networks (and eventually general AI), might collectively re...