Spectrum Policy is Too Politicized
Congress is struggling to find a way forward on spectrum licensing policy. The spectrum pipeline is bare, the FCC’s auction authority is on life support, and the national spectrum strategy NTIA and the FCC are supposed to be developing is a long way from completion.
While Congress stands still, markets for wireless-enabled services continue to grow, generating trillions per decade in economic value from licensed networks alone. On the unlicensed side, Bluetooth and Wi-Fi are ubiquitous, even if their economic value is harder to quantify; these services likely contribute hundreds of billions if not trillions as well.
The last FCC was very generous to Wi-Fi: Its 6 GHz order provided 1.2GHz, tripling the amount of unlicensed spectrum Wi-Fi can use in the US. The regulator has been much more parsimonious with licensed spectrum. In the 5G and 6G C-band sweet spot, the FCC relied on voluntary surrender of 300 MHz by incumbent GEO satellite operators moving into the slot between the C-Band and the radio altimeter band.
What Comes Next?
The incentive auction freeing up the C-Band generated $81 billion all by itself, making it the most expensive (in terms of dollars per Mhz) spectrum auction ever. Demand for licensed spectrum is clearly high. Meanwhile, consumers are moving into the 6 GHz Wi-Fi band at a snail’s pace as it provides little noticeable improvement to non-gamers per our testing. Device manufacturers are shifting to 6 GHz chipsets, but consumers are not upgrading routers.
Notwithstanding the high demand for licensed spectrum and the imbalance in recent allocations, some in Congress (chiefly on the Democratic side) would like spectrum auctions to simply stop. This faction responds to demands from incumbents with heavy investments in legacy tech (such as DoD,) cable companies, populist policy wags, and visionary technologists who see a future in which radio interference is a solved problem.
There’s a lot to argue about, and all stakeholders have interests, perceived or legitimate, that will have to be resolved for the nation as a whole to progress. It’s important for lawmakers to enter this debate with an interest in learning as well as freedom from emotional baggage that generates bias against particular stakeholders.
How about a Little Sis-Boom-Bah?
As veteran telecom reporter John Hendel tells it, bias in favor of Wi-Fi plays a big part of the in the current spectrum debate (as if Wi-Fi were in danger):
As a policy debate, the fight to prioritize Wi-Fi and other unlicensed use of the airwaves is central to today’s spectrum arguments in Washington, and their outcome could dictate the next several years of U.S. tech policy. Congress is negotiating a package of spectrum legislation eyeing these issues, still largely in flux but pegged to the reauthorization of FCC spectrum powers set to expire in March.
And this debate is likely to be central to the Biden administration’s promised National Spectrum Strategy, which could come this year, and set goals for how airwaves get into the hands of the private sector.
The feeling that Wi-Fi is the mother of innovation reflects the relentless and very successful public relations campaign executed by Big Tech and its minions. Hendel summarizes the campaign’s key claim:
The whole thing shakes out — broadly — into an argument between the big telecoms that want more spectrum to carry their cell signals, and the consumer-tech and cable companies who want to be sure there’s a wide-open playing field for device innovation.
[my emphasis] So Big Telecom wants the airwaves for their very own cell signals, but scrappy little Wi-Fi carries the entire innovation mantle for all of us. This framing ignores or devalues practically every app on our smartphones.
Clearly, no one technology has a monopoly on innovation. In the real world we have networks inside our homes and offices and outside them. Wi-Fi can’t do what 5/6G can do outdoors even with all the help enthusiastic lawmakers can give it; we need both licensed and unlicensed spectrum to enjoy the full scope of innovation. How can anyone in their right mind think otherwise?
Where Does Internet Anti-Telecom Bias Come From?
The bias against telecom in the Internet space has a long history. The oldest story about the battle between Internet innovation and the phone company is the Internet’s own creation myth.
The myth pits computer people – the so-called “Netheads” – against telecom people, the “Bellheads.” It comes from a demo of Internet precursor ARPANET given by Ethernet inventor Bob Metcalfe to AT&T executives. It didn’t go well:
“I’m sitting at a terminal,” recalls Metcalfe, “this graduate student with a huge red bushy beard, giving a tour of this network to 10 executives from AT&T, all of whom were wearing pinstripe suits…and in the middle of my demo – for the one time in the whole three days – the system crashed. And I looked up. And they…were happy…that it crashed. They were smiling.” Metcalfe is still incredulous. “This was my life’s work. My crusade. And these guys were happy that it didn’t work.”
The takeaway? “I saw that there are people who will connive against innovation,” he says. “They’re hostile to it. And that has shaped my behavior ever since.” Metcalfe’s is a world of good guys and bad guys, and since that day in 1972, noncompetitive telcos have been planted firmly in the latter category.
Practically every time I’ve met an Internet person of Metcalfe’s generation the conversation veers toward the wickedness of “the powers that be” within seconds. When I read “a wide-open playing field for device innovation” Metcalfe’s tale always comes to mind.
Good Guys and Bad Guys
Even though the AT&T of 1972 isn’t today’s phone company and the ARPANET of that era predates TCP/IP, the tendency to label one group good and the other bad will always be the lazy person’s route to resolving complex problems.
This naïve framing drove the net neutrality controversy created by law professor Tim Wu in 2002. At a Silicon Flatirons event, Wu argued that broadband providers needed to be heavily regulated else they prevent Silicon Valley entrepreneurs from achieving their full potential for greatness.
The title of Wu’s paper “Network Neutrality, Broadband Discrimination” tells you who the good guys and bad guys are. Twenty years later we still don’t have a network neutrality law in the US and the broadband plant is the least of the Internet’s worries. The framing is naïve.
The Dualist Narrative in Spectrum Policy
While none of Wu’s dire predictions panned out, he is nevertheless a highly respected figure in Internet policy, publishing several books, securing tenure at Columbia Law, and serving as the White House chief tech policy guru. The politicization of spectrum policy appears to be one of his legacies.
It doesn’t have to be this way. Spectrum licensing enables a range of services that unlicensed can’t provide, such as high performance wide area connectivity, GPS, satellite broadband, and microwave backhaul. And unlicensed spectrum provides a range of services that don’t benefit from licensing, such as Bluetooth headsets and Wi-Fi appliances.
In a rational world, licensed and unlicensed must coexist, as they do when your Wi-Fi session at Starbucks determines its location via GPS or hands off its data to a microwave backhaul link. Getting from either/or to both/and has to be among the the key goals of the national spectrum strategy. If either one must be prioritized, we need to look at the objective benefits.
Efficiency is Enhanced by “Licensed First”
In the beginning, the wireless part of the Internet was designed to work around licensed spectrum allocations. The first example was the Bay Area Packet Radio network debuted by Don Nielson’s team in 1976. It sought to avoid interference with licensed systems by using spread spectrum technology.
Wi-Fi originally followed this path, using spread spectrum in a junk spectrum band used by consumer microwave ovens. Wi-Fi was designed to tolerate interference on the reception side and to minimize it on the transmission side by using low power over short distances. Licensed systems generally use higher power in order to cover miles, while Wi-Fi only promises to reach 200-300 feet.
CBRS extends the concept of licensed first by offering licenses to new spectrum when and where they are needed, while allowing unlicensed systems to scavenge unclaimed spectrum for free. Unlicensed simply finds a band (or combination of bands) of spectrum sufficient to get the job done that isn’t being used by a license holder. Where the market demands licensed services, they can also be provided.
This system of priorities promotes efficiency and flexibility. Under the best of circumstances, Wi-Fi is only able to utilize spectrum half the time because the protocol waits for the network to be idle before transmitting. Licensed systems manage spectrum by schedule, and can use 95% of available capacity.
How Innovation Really Works in Broadband
The history of innovation in the networking space is rife with examples of great Nethead ideas that didn’t come to fruition until Bellheads took care of the deeply technical engineering details. The original standard for Metcalfe’s Ethernet – the Blue Book devised by a committee of computer people from DEC, Intel, and Xerox – called for attaching every computer to a shared coaxial cable and implementing an elaborate protocol (“carrier-sensing multiple access with collision detection and randomized binary exponential backoff“) to determine who got to transmit.
The cost of the coaxial cable and the electronics attached to it made potential users wary of installing Blue Book Ethernet. The IEEE 802.3 standards body accepted it anyway after adding its own set of options. When cost became a barrier to adoption, 802.3 chartered a task force to make it cheaper [I was the vice chair of this task force, known as the StarLAN task force within 802.3 and later a Wi-Fi contributor.]
The key to lowering the cost of installation was to ditch the shared cable and the delicate protocol in favor of a shared piece of electronics – a hub or switch – with individual telephone-style wires between the hub and the computer. This approach allowed multiple kinds of cabling, multiple speeds, prioritization, and fault isolation. Hub and spoke, as it’s called, is not just cheaper, it’s better and more scalable than a shared cable. Most of the ideas for this new approach to Ethernet came from people working for AT&T divisions or marketing to them.
Something similar happened with Wi-Fi. Some of the people from the StarLAN task force contributed ideas for the organization of the 802.11 network and its access protocols that built on StarLAN work. This is why we have Wi-Fi routers instead of a Blue Book-like network of peers.
Who Do You Trust?
When I say the early trust of Internet-native businesses and distrust of telecoms was naïve, I’m comparing the size and reputation of Big Tech firms today vs. their charmed status when the network neutrality drama started in the early naughts. In those days, Google, Yahoo, Netflix, and Amazon were scrappy little startups and Twitter and Facebook didn’t exist.
The telecom world was mainly highly concentrated wireline telephone and broadband service and cellular was a novelty. It was rational to fear the large monopolies and to promote the development of the innovators in competitive markets. But the tide has turned dramatically since 2002.
The Black Hole of Unlicensed
Once the FCC designates a spectrum band for unlicensed use there’s no turning back. Unlicensed means anyone can use it for anything at any time with little regard for infrastructure investment, reliability, or efficiency. Flexible use licenses are just the opposite, incentivizing their holders to do more with less.
It’s no wonder that the Wi-Fi industry’s next step after the FCC’s gift of 1200 MHz of spectrum was to ask for even more spectrum in the 7GHz band. We’re in the early days of 6 GHz Wi-Fi and its success remains to be proved. There are more critical, demonstrated needs for 7 and 12 GHz. While 6 GHz Wi-Fi is utterly uncongested and little used, we do see congestion on residential 5G networks.
T-Mobile essentially pioneered this service in the US, and they’ve suspended subscriptions in some areas because of network overload. If spectrum policy tracked demand, residential 5G would be the priority. And Wi-Fi is a non-starter for residential broadband in normal suburban neighborhoods even if it works great on 15 acre hobby ranches in central Texas.
Set Network Innovation Free
From the standpoint of increased network efficiency and resilience, Wi-Fi is a weak player. For all its cost and complexity, licensing does a better job of matching supply and demand than the model that takes licenses off the table in order to subsidize inefficient networks. Flexible use licenses perform extremely well.
Wi-Fi and licensed 3GPP (5G and 6G) have a symbiotic relationship in which advances in signal processing, modulation, encoding, and spectrum reuse are shared between the two. Chipset producers benefit from learnings from the other side the licensing divide, the rapid revision of Wi-Fi standards, and the perfection of novel concepts in licensed systems in which devices and infrastructure are well integrated.
Starving one side of this divide to nurture the other can only produce dire results for innovation in network design and technology development. The worst thing the US can do to spectrum policy would be to import political partisanship, grievance, and score-settling into the process of policy development. Congress needs to rise above these divisive tendencies to craft sound spectrum policy.
Spectrum policy also needs to be guided by the realities of network engineering rather than the desires of network incumbents to protect legacy business models, legacy reputations, and legacy technologies from upstarts. Spectrum policy need not be an exercise in exclusion, it must focus on enhancing the vibrant exchange of ideas among all players in the innovation ecosystem.