Sign up for tech policy updates
The latest on technology policy from AEI, published daily.
We can’t see the future: Prioritizing innovation on the 5G internet
On April 17, Paul Schroeder sat before the House Energy and Commerce Committee with an array of legislators looking on. But Schroeder couldn’t see his inquisitors. Schroeder is blind, and he was demonstrating an ingenious new system that exploits mobile broadband to help the seeing-impaired navigate the world around them.
Schroder’s testimony also exposed a key weakness in the never-ending effort to impose heavier regulation on the internet, often referred to as “net neutrality.” The now-defunct Title II Order, issued by the Federal Communications Commission (FCC) in 2015, banned “paid prioritization,” or services that speed bits to their destination, which might be especially useful for real-time applications — such as Schroeder’s super goggles. But paid prioritization has been a sticking point in the debate for years. Understanding what paid prioritization is and isn’t may thus be important to a Congressional compromise that could settle the net neutrality debate once and for all.
At the hearing, Schroeder wore special Aira Technology eyeglasses sporting an embedded camera, microphone, sensors, and wireless connectivity. On the other end of the line was a representative in San Diego, who acted as Schroeder’s eyes. She was watching and listening — and describing to Schroeder — the real-time visual feed of his surroundings. “The ranking member is waving to you,” she told him.
Imagine how much easier it would be for the blind to maneuver through an airport, a crowded event, or a hiking trail with real-time guidance. Imagine how much richer the experience with a partner describing the colors, vistas, obstacles, and people.
Unlike email, texts, webpages, or even one-way streaming video, however, Aira’s seeing-eye service can tolerate almost no delay in transmission. “Aira has an interesting case to make regarding prioritization,” Schroeder testified. “We need to send video upstream, and as you know, often upload speeds don’t match download speeds. Our critical case to make is that our service can’t work if we don’t have priority, low-latency access.” In real-time scenarios, delay can ruin the experience or even cause harmful accidents. “If somebody is out and about moving,” Schroeder said, “they really do need that instant video feedback that the agent can then provide, but that video is moving in the opposite direction of what we usually talk about.”
The original goal of net neutrality proponents was to ensure all bits are treated equally, that no content gets an unfair advantage over other content. “No paid prioritization” thus became a key pillar of the movement. But the real objective was to avoid a scenario where an internet service provider (ISP) favored its own content or disfavored a rival’s content. What should have been a competition question, addressed by existing antitrust and Federal Trade Commission rules, instead was turned into a broad ban on technologies and services that actually might be very useful and good for consumers.
“No paid prioritization” became a political rallying cry that almost no one understood. It was presented as a way to ban “fast lanes” and “slow lanes,” as if Google and Amazon would pay for speedy service and Main Street websites would be stuck in traffic. But guess what? Google and Amazon had already created their own fast lanes, known as content delivery networks (CDNs), 15 years ago. And millions of main street websites use third party CDNs, such as Akamai, to store content closer and thus speed content to their users. Early versions of net neutrality legislation would have banned CDNs, already in use 15 years ago. Thankfully Congress didn’t enact the proposals.
In a world without heavy-handed rules, the US internet boomed spectacularly. Network capacity and infrastructure improvements, such as CDNs, set off an explosion of apps, content, and new connected devices. The US generated more data traffic per user and per capita than any nation on earth. In our need-it-now world, no lane is ever fast enough. But the US built the fastest lanes in the world. Meanwhile, lots of other factors, such as proliferating pop-up ads, were doing far more to slow webpage downloads and degrade the user experience than any ISP practice.
The FCC’s 2015 Order had five chief components: a transparency rule, which said ISPs have to disclose their network policies; three bright line rules against blocking, throttling, and paid priority; and a general conduct rule. A modified transparency rule survived the 2017 Restoring Internet Freedom Order (RIFO). The expansive general conduct rule was the FCC giving itself almost unlimited regulatory power over undefined future technology and behavior, and RIFO dispatched it quickly. No blocking and no throttling, on the other hand, are good rules of thumb, and almost everyone across the internet world agrees. Although the FCC found these rules beyond its legal authority, nearly everyone supports legislation to prohibit blocking and throttling, which would seem to take the “slow lane” fear off the table. No network can degrade someone’s lawful traffic based on its content or origin. Paid priority thus remains one of the last disagreements to be hashed out.
Sign up for The Ledger newsletter
And here’s where a better understanding is needed. In some sense, the internet equivalents of next day air and three day ground delivery already exist. For a price, CDNs store content closer to the end points. For a price, higher bandwidth links can transmit more data per unit of time. Different products for different needs at varied price points.
For real-time services, however, cached data or more capacity aren’t good enough. I cannot “store” a video conference closer to the user, nor is more capacity always the answer. I don’t need a larger box for a time-sensitive letter. Sometimes, real-time data requires prioritization. Like Paul Schroeder’s seeing-eye glasses. Or collaborative online gaming. Or augmented reality. Or, soon, autonomous vehicles driving on busy roads. And remote-controlled robotics for industry or medicine. Creative applications like these will only become more common with the advent of 5G wireless, as we connect more devices and real-time video continues to explode.
The list will go on, and we don’t know exactly which applications will require priority service. But it’s hard to argue that we need a preemptive ban.
One argument for a ban says that networking is zero-sum: that prioritizing one packet automatically degrades another packet. But network economics are not zero-sum. The offering of overnight parcel service, for example, does not slow the delivery of the three-day package. In fact, the premium offering can improve the infrastructure and economics of the firm or network so that the basic three-day service is supplied at a much lower price. Everyone is better off. In the same way, premium internet services make other broadband options less expensive.
Advocates of a priority ban correctly point out that the 2015 Order had an exception for specialized services, which are sold separately from basic internet access service (BIAS). The Aira smart-glasses service might fit this exception, for example, and so might connected cars. But the exception was subject to the discretion of the FCC, which could have ruled that specialized priority offerings were just a way to avoid the BIAS designation. The exception was far better than a ban on all prioritization across the internet but still vulnerable to bureaucratic tampering.
It’s even more cumbersome than that, however. What if a consumer wants to purchase a one-time boost for gaming with friends on a Friday night? Or use an augmented reality app for work? Or try out a new real-time service without signing up for a separate subscription? Forcing ISPs or app developers to offer specialized services apart from the basic broadband, mobile, or app could hurt consumers by forcing them to buy an extra product instead of an integrated feature or zero-rated app within their existing subscription.
Any new legislation should thus of course preserve an exception for specialized priority services. The exception should be the floor or starting point. Even better, however, it would acknowledge that prioritization will be a bigger (although unknowable) part of the 5G wireless internet than of previous generations. It would acknowledge that anti-competitive practices are already being policed at the FTC, which is on the lookout for discriminatory anti-consumer behavior.