Effective and Efficient Wireless Networks
Commerce Secretary Gina Raimondo made a sage observation while kicking off Monday’s NTIA Spectrum Symposium:
Without effective, efficient use of spectrum we won’t have innovation. We need to free up more spectrum to unleash innovation. We rely on spectrum at NTIA for the life-saving work of FirstNet…there’s no FirstNet without spectrum, there’s no innovation without spectrum.
I’m not convinced that NTIA staff are aligned with their boss. Instead of focusing on efficiency, the symposium largely endorsed popular but relatively inefficient spectrum access systems, giving special deference to the highly random grab bag of ideas that emerged from a recent Aspen Institute policy roundtable of lobbyists, lawyers, and regulators.
NTIA needs a course correction if it is to contribute in a constructive way to the national spectrum strategy DC policy mavens are kicking around. Secretary Raimondo got it just about right, but her employees aren’t fully on board.
Two and a Half Ways to Allocate Spectrum
The core question in government spectrum policy is whether resources should be managed by the marketplace or by government fiat. Markets produce messy results, so the US and most other countries first tried the command-and-control method.
This resulted in multiple instances of protecting incumbents from competition and the strange decision to relegate television to the airwaves and telephone to wires. The US was slow to adopt cellphones because our regulators saw them as novelties for the rich for far too long.
The FCC finally landed on spectrum auctions for licenses and general purpose unlicensed authorizations for low value uses in the 1990s. They supplemented this two track system with a fledgling hybrid called Citizens Band Radio Service (CBRS) in 2015.
General Purpose Spectrum Licenses
Because spectrum auctions were so late to the game, they have to work around pre-existing allocations made piece-by-piece with little sense of a big picture. Each license therefore inherits restrictions on transmission power made for the original system, with whatever degree of liberalization deemed safe by the FCC.
Allocation by license is therefore re-allocation in nearly all cases. Incumbents in adjacent bands are often able to block new entrants by claiming interference even when such neighbors are out of specification, as is the case with GPS users and aviation altimeters.
Licensed spectrum is more efficient than other access modes because it shares information best. In a cellular system, base station code knows how many users are active and how much data they need to send and receive. Hence, these systems can utilize 98% of their allocations, the epitome of efficiency.
Unlicensed Authorization by Rule
Unlicensed spectrum is governed by rules enforced by the FCC through certification. Wi-Fi vendors simply have to prove that their products conform to agency limits on power and frequency to get permission to operate.
While licensed devices share information with the network by design, unlicensed devices play on a very uncertain field. In the absence of a control point able to coordinate every access to the air among multiple users, unlicensed devices operate mainly on hope, luck, and dead air.
Wi-Fi devices will only transmit after sensing dead air for a period of time corresponding to the latency of the local network from one end to the other. In practice, these networks tend to idle 50% or more of the time. This is low efficiency.
CBRS is a mashup of licensed, unlicensed, and pre-emption. I consider it progeny of IEEE 802.11y, a super Wi-Fi mode designed to use Naval RADAR frequencies in areas where the Navy doesn’t use them.
CBRS base stations (Citizens Broadband Radio Service Devices or CBSDs) assert control over user stations such that civilian operation can be disabled or controlled. The Navy can use the entire block of spectrum whenever it wants. When the Navy is not active, roughly half of the band can be used by license holders with the rest available to the public on a Wi-Fi-like basis.
When license holders are inactive, their spectrum falls into the public pool. CBRS is appealing to wireless ISPs because licenses are inexpensive, covering small slices in single county areas. CBRS is experimental, a potential solution to a number of networking problems at the margins of common use.
Who Wins on “Effective and Efficient?”
Only flexible use licenses provide peak effectiveness and efficiency. Unlicensed is inexpensive but it loses on effectiveness and efficiency by virtue of its complete lack of interference protection and coordination.
CBRS is experimental so it’s unclear that it will ever be suitable for applications other than small ISPs and some proprietary network applications. A variation with even smaller licenses might prove useful for campus and factory networks running forms of LTE and 5G.
But the Spectrum Access System is a single point of failure and good behavior can’t always be assumed on the part of end user devices. CBRS maximizes flexibility over efficiency, hence its appeal to regulators.
National Spectrum Strategy
The US is clearly headed toward a National Spectrum Strategy and a ten year plan for spectrum. But we still lack a sound technical foundation, as we can see from the Aspen report. Consider this comment on page 28 of the report:
Any spectrum strategy should focus on the success of wireless technologies in the short term, and maintain flexibility for technological advances in the future. It is important to set ambitious but achievable goals that are rooted in sound interagency processes and, to the extent possible, anchored in specific spectrum bands. [emphasis added]
Why would we not root our ten year plan in sound technology rather than on the ability of government actors to work together? The purpose of our plan should be promoting innovation, not simply protecting bureaucratic fiefdoms.
Review Government Systems with Zero-based Spectrum Budgeting
The NTIA symposium did a good job of bringing lesser known agencies to the table: NASA, NOAA, and FAA were all represented, as was the all-powerful military. But each of these agencies showed that it regards its present mode of spectrum use to be set in stone.
A plan needs to be driven by technologists and economists as well as lobbyists and interagency process, and agencies need to be motivated to rethink their current spectrum practices. Many government systems can be replaced by modern upgrades with zero incremental spectrum footprint above the commercial and private systems on which the highly productive civilian sector depends. Look at FirstNet.
Government spectrum systems often hail from an era where operating their own wireless gear was their only option. That is no longer the case, obviously. Perhaps the next step is a challenge process for replacements to out-of-date spectrum systems.