Broadband in the Lockdown Era

All of us who can are working, shopping, and enjoying entertainment from home now. My quarantine is especially strict because I’m in a high-risk group.

Store shelves are still stripped of toilet paper, sanitizers, bread, flour, and yeast, but kale stocks are holding strong. Broadband is a paradox because it’s also holding strong while utilization is as great every day as it usually is during the Christmas season.

Sadly, some policy advocates are seeking to exploit the lock-down to give a boost to their hobby horses. While it’s an axiom of politics that no crisis should go to waste, the exploitation is still unseemly. Some people are having trouble with their Wi-Fi networks, but the Internet is outperforming physical goods networks by a huge margin.

Changing Patterns of Utilization

The most intriguing observation I’ve seen is about the Internet’s versatility. During normal times, we use our home broadband connections to communicate with friends and colleagues by text, to browse the web, and to watch video from streaming services.

But some alternate forms of communication are increasing, sometimes quite radically. Video conferencing consumes more upstream capacity than streaming, because it’s bidirectional, so it’s not surprising that NCTA reports upstream traffic up by 28% in the last month.

Normally, BitTorrent and VPNs account for most of the upstream traffic on the US Internet – about 16% combined. In other parts of the world, BitTorrent is an even bigger factor, accounting for 44% in Europe, Middle East, and Africa (EMEA).

Weak Links

Reports of sub-par Internet in EMEA suggest that upstream is more likely to be the problem than downstream. Ookla data show Europe’s broadband has less capacity than North America’s, so there’s not that much margin for absorbing more traffic.

Combining Europe’s higher use of BitTorrent with its lower overall capacity and the fact that broadband upstream capacity is less then downstream everywhere suggests that Europe is a lot closer to the cliff than we are.

This makes the EU’s order for video streaming services to reduce their resolution look suspect. In general, video streaming services dynamically adjust their bit rates and resolution to available capacity, so there’s not any reason to slow these services down to SD resolution all the time.

Even BitTorrent does that, using the LEDBAT protocol. The main conclusion I would draw about Europe’s broadband today is that their regulators are pushing the wrong levers because they lack a good understanding of the system they’re regulating.

Imagining the Worst

If there were to be a scenario where the Internet stops working well enough for all common apps to hum along, it could likely come about from BitTorrent, Skype, and Zoom saturating the uplink. This is unlikely to happen because all of these apps are dynamic consumers of bandwidth.

What you would see would be more low resolution video, more of the time, the exact situation that Europe’s regulars want to establish as the new norm. In the US, Internet overload would be the condition in which our networks behave like Europe’s.

We’re nowhere close to that because we’ve only see fairly modest increases in consumption relative to baseline growth. The Internet always carries more data in any given year than it did the year before, so adding capacity is part of the ISPs’ daily regimen, like taking a vitamin.

The fact that we’ve seen as much growth in the last month as we see in the typical year is therefore not alarming. If the condition continues for several more months strange things would start to happen, but we’re not there yet.

Legitimate Problems Have Always Been Hard

While fears of imminent Internet collapse aren’t based on reality, three related issues are worthy of attention: the rural divide, poverty, and Internet hesitancy. We don’t know exactly how extensive broadband networks are in rural communities because precise mapping is expensive. We do know that claims of poor build-outs in urban and suburban communities are essentially nonsense, however.

We know that many people access the Internet exclusively through mobile devices because they don’t have the money for a fixed-location broadband service at home, they don’t want to spend their last $40/month for broadband, or they don’t have a fixed address. This problem is a manifestation of poverty, something that still exists in the US long after President Lyndon Johnson’s War on Poverty was supposed to fix it. It won’t be solved anytime soon.

And we know that a non-trivial portion of the population sees no value at all in using the Internet, or at least not enough value to outweigh the negatives. We’re not going to eliminate identity theft, fake news, and generally rude behavior on the Internet anytime soon either.

Where the Lockdown is Helping

The lockdown has certainly changed the risk/reward calculation by providing concrete, compelling reasons for being connected. Working from home during the lockdown enables people to earn paychecks who wouldn’t otherwise be able.

Being connected at home also means the kids can attend school and the parents can keep up-to-date on the latest pandemic news. It also means that people are able to learn about things they need to know by watching YouTube.

For the first time in history, Sandvine sees more YouTube traffic than Netflix traffic, and much of that has to be educational videos. This is a mixed blessing because so much YouTube content is dubious, but overall it’s probably somewhat good.

An Issue Worth Exploiting

Internet inclusion – getting everyone connected in some way – has been one of the more vexing Internet policy problems since the 1990s. The Clinton Administration did some good work on the problem in its day, with its E-Rate program, but no other administration has been equally enthusiastic.

Now that we understand the Internet is essential during pandemics, perhaps that will change. And this is not going to be the last pandemic.

I applaud the advocates who are working to ensure that we end the so-called “homework gap”, even if they step outside their lanes on occasion. Ensuring that every school-aged child has a computer at home and the ability to access the Internet is important.

Three Prongs of Connected Kids

This problem is fundamentally a combination of rural broadband expansion, subsidies for computers and connectivity, and convincing parents to support the use of computers in the home. Programs that fail to include all three prongs will fail, as they have for the past 30 years.

At least part of this problem will depend on making the necessary enhancements to the Internet architecture that will enable us to provide a kid-safe Internet experience. This may be the hardest part, actually, because it’s the only part that’s not simply a matter of money.

Schools can start to treat laptops like textbooks, simply issuing one to every student who doesn’t already have one. That leaves the mapping problem to be solved, and I suspect the Census Bureau has a role to play. We’ll take that up in a subsequent post.