Download The Report
New Policies to Serve New Users of Spectrum
The second half of the Aspen Institute Roundtable on Spectrum Policy consisted of three breakout groups that gave focused consideration to regulatory challenges in three areas:
Internet of Things and Drones: New Users and Uses
Working Group A considered how drones and technologies based on the Internet of Things (IoT) are creating novel uses of spectrum, which in turn require a wholesale assessment of how regulatory policies should be structured. How should the Internet of Things fit into the whole universe of policies, especially telecommunications?
There are some bodies who have or are currently investigating the options, including an historic IoT working group and a current working group studying drones and satellites within the FCC’s Technological Advisory Council. The General Accountability Office is currently completing a report on IoT to attempt to understand its general dynamics and situate IoT issues within existing policy frameworks. It helps to understand that IoT technologies span a highly diverse array of fields and applications, so the very term IoT may suggest generic similarities among the technologies that are tenuous at best. For example, IoT includes everything from drones and autonomous vehicles to industrial sensor systems, home security systems, environmental sensors and wearable sensors. They are likely to have different performance demands and therefore use different communications networks; have different scales, degrees of independence and control regimes; pose different safety risks; and function with different market economics.
So how, then, to regulate IoT technologies? Working Group A came to a general consensus that applications that do not affect public safety and are not related to critical infrastructure—i.e., smart appliances, home security systems, wearable sensors, medical devices, location cuffs—can be accommodated under current licensed and unlicensed spectrum standards. This conclusion also applies to the anticipated evolution of IoT technologies.
However, for IoT applications that affect the safety of life and critical infrastructure—e.g., autonomous vehicles, drones, sensors for chemical, biological or nuclear defense, industrial infrastructure—a different set of policy concerns should apply. While licensed spectrum can still be acceptably used, the group concluded that there must be a rigorous prioritization of service for these critical, safety-related applications. For example, an unmanned drone—which can sometimes be as large as a conventional cargo aircraft—needs to communicate its control information on an on-demand basis. Such IoT uses can potentially be accommodated on high, mid- and low-band spectrum, but require adequate coverage and capacity.
Working Group A identified key technical considerations that come into play when making policy for IoT technologies. These include:
In setting policies, legislators and regulators also should consider:
Working Group A identified a different set of concerns for drones, along with appropriate policies for regulating their usage. There are three basic classes of drones: hobbyist drones that are 55 pounds or less, fly at less than 400 feet and can be controlled through line of sight transmissions, light commercial drones that are very similar, but with greater capacities such as control without line-of-sight connections; and large aircraft up to and including Boeing 767 size aircraft, run by cargo and logistics delivery companies. Many industry groups are working on these issues. The FAA and NASA are currently conducting trials, but there are many concerns to identify and rules to formulate.
Among the concerns: Users must have real-time control, which requires secure spectrum access to ensure reliability and safety. Safety is a significant concern, especially to protect against hacking and spoofing, or other variables that could jeopardize control (flight factors, physical payload).
Payload variables can also affect the safety of drones. The weight cannot be above certain limits, and broadband availability must be sufficient to avoid interference problems in transmissions. Variables such as real-time data analysis versus store-and-analyze functions must be taken into consideration. The size and weight of a drone can matter, also, along with its range of travel.
The Panel identified the most likely applications of drones in small and large categories, with “small” being less than 55 pounds—the current drone weight limit specified by the Federal Aviation Administration—and “large” being anything from 56 pounds to the size of cargo planes. The most common drone uses include:
In each of these cases, the panel concluded that drones can use existing (and enhanced) licensed and unlicensed spectrum and standards for payload communications and line of sight control. However, when the drone is not controllable through a line of sight, it must have “dedicated spectrum” on a licensed, time-based standard, which carriers could provide through various prioritization and preemption techniques, such as SAS [spectrum access system]. The drone operator in such cases must pay for service from an appropriate service provider.
The group concluded that IoT technologies generally need access to the entire stack of spectrum—high, middle and low spectrum, all together—to cover the range of services required ranging from latency sensitive applications and to applications requiring in-building cover and therefore the ability to penetrate through walls and so forth. With the entire stack, policy could be developed to provide prioritization for certain types of mission-critical, safety needs, and perhaps for certain commercial uses as well. An unresolved question is whether entities like Federal Express or UPS could reliably use spectrum to exercise remote control drones the size of cargo aircraft.
It is envisioned that carriers could provide licensed “dedicated spectrum” to drones on demand, at the times needed, through a sophisticated prioritization and preemption system (which the current 4G/LTE standards supports). If large drones needing such spectrum never took off in flight, for example, the spectrum would not be affected, and would just be used on a cellular basis. SAS-like structures might be used if needed for high-priority purposes, but would not necessarily serve as a general default. While SAS is regarded as a very good model for sharing spectrum, participants noted that the technology, at least in 4G/LTE wireless contexts must be proven first. (This raised the question if such preferential treatment would amount to a violation of net neutrality rules. Perhaps not, if the offer were made to everyone.)
The higher-level spectrum policy framework should look like this, the panel concluded: Applications that implicate safety of life and critical infrastructure—e.g., heavyweight commercial drones or autonomous cars―should use licensed spectrum and carriers should modulate usage through a prioritization and preemption model to ensure that time-sensitive, safety-related communications can work reliably. All other communications should be supportable by either licensed or unlicensed spectrum as appropriate for the specific application.
Jonathan Chaplin of New Street Research summed up the sentiments of the panel, saying, “We favor spectrum use policy that gets us away from dedicated spectrum for dedicated users for dedicated purposes. We want a spectrum allocation framework that maximizes for efficiency and utilization, to make the best use of a scarce resource. But we need sophisticated prioritization processes for critical safety issues and non-line-of-sight drones.” This approach would shift control to carriers, who have the capacity to exercise this kind of management. Carriers might not necessarily set the rules, but they would administer them.
One concern was the contingencies of shared spectrum for drones. Carl Povelites, Assistant Vice President of Public Policy for AT&T Services, worried that “innovation and investment would be chilled by the uncertainties about what you can and cannot do with the spectrum, because people could be preempted by prioritized uses.” That is why “exclusive flexible spectrum” may be the best approach here, explained Dennis Roberson of the Illinois Institute of Technology. “Carriers would control the spectrum directly. That would be lost with shared spectrum, which in turn could affect investment (because control of the spectrum would be less predictable).”
Other complications arise, however: How would a carrier or regulators prioritize a lifesaving drone in the air versus autonomous cars on the ground? The group conceded that the economics of the preemption/prioritization scheme are also uncertain and perhaps too complicated and expensive. Can SAS technology be made more versatile and reliable? And how can the system prevent bad actors from spoofing transmission signals, such as signaling “I’m an aircraft.” The panel emphasized that implementation and technology development over time will be critical, but that this policy framework remains theoretically appealing and practical for drones.
Working Group A consisted of Dennis Roberson (coordinator), Charla Rath, Marjory Blumenthal, Paula Boyd, Jonathan Chaplin, Emilie de Lozier, Nicol Turner-Lee and David Bollier.
Sorting the Uses: Licensed, Unlicensed and Shared
Working Group B was charged with coming up with new policy recommendations based on expected spectrum needs in the coming decades, with methods for going beyond those needs if necessary. Of course, it is difficult to predict future spectrum uses and needs; uncertainty is certain. As Larry Downes, a member of Working Group B and Project Director of the Georgetown Center for Business and Public Policy, put it, “We generally get less change than we expect in five years, but more than we expect in ten years.” Change happens slowly, then suddenly and unexpectedly.
Given these caveats, Working Group B expects that in five to ten years, interference issues will become less of a deterrent to flexible licensing, making unlicensed bands more responsive to new users and increasing the ability of secondary markets to reallocate licensed bands as needs change. That is because even as unpredictable new apps create substantial demand for spectrum already allocated, transmitters and receivers will become much better both at avoiding interference with nearby frequencies and in dynamically changing bands when a user with higher priority appears.
The group presented three recommendations:
Recommendation No. 1: Adopt a flexible use paradigm for new applications.
Public policy should promote flexible use for both licensed and unlicensed spectrum, as recommended in item 5.1 of the National Broadband Plan. Flexible service rules should be limited solely by interference tolerances—a “harm claims threshold” with other proximity bands (licensed) and within bands (unlicensed). This recommendation anticipates that both transmitters and receivers will increasingly minimize interference.
This recommendation harkens back to the work of Ronald H. Coase, whose 1959 essay, “The Federal Communications Commission,” argued, “There is no analytic difference between the problem of interference between operators on a single frequency and that of interference between operators on adjacent frequencies. The latter problem, like the former, can be solved by delimiting the right of operators to transmit signals which interference, or might potentially interfere, with those of others. Once this is done, it can be left to market transactions to bring an optimum utilization of rights.” Downes elaborated: “If we define interference as a property right, then private negotiation can better resolve issues without FCC intervention or detailed regulation or allocations to carve up spectrum bands.”
This approach has never really been tried, however, noted Downes, because of political involvement in the policymaking process, rent-seeking behavior by incumbents, and agency incentives to consolidate power rather than allow markets to resolve conflicts. For example, relying on market resolution with credible government enforcement as a backstop did not work (at least in the initial rounds of negotiation) in the dispute between LightSquared (now Ligado Networks) and various GPS device manufacturers over potential interference with GPS signals (where a private solution would involve payments to retire a legacy technology), the group noted. In addition, the tort system, which Coase saw as the more efficient solution, is an increasingly expensive and uncertain venue for resolving issues.
Working Group B believes that the best solution to interference disputes requires an institutional design that favors negotiated rulemaking and a single vote at a federal agency (the top executive or bureau chief). Arbitration can be used as a stick to prod a resolution, or mediation as a carrot. In any case, the group proposed that the burden of showing harmful interference should shift from the entrant to the incumbent user of spectrum (an adjacent user for licensed spectrum, an existing user for unlicensed spectrum).
The group acknowledged that this approach may not work if the parties are in asymmetrical power or information relationships—e.g., in which the newer user is smaller, poorer and lacks relevant data on the technical characteristics of the incumbent’s equipment. Other issues to pay attention to: Is the arbitrator an expert? How should incumbents be incentivized to accept the paradigm? (VIA, the Voluntary Incentive Auction, is perhaps a good example.) And should there be different incentives for federal incumbents (versus a simple fiat)?
Recommendation No. 2: Adopt a new model to free up capacity and move inefficient assignments to the flexible use paradigm.
The group recommended that the FCC should apply both carrots and sticks to get obsolete technologies and incumbents to vacate existing spectrum as quickly as possible so bands can be reallocated for new uses and under flexible license that could subsequently be reassigned in the secondary markets. In the interim, before the obsolete uses are retired, obsolete spectrum assignments under inflexible licenses should be regarded as “transitional bands” and shifted to tiered access (“shared”) with new users (the term “shared spectrum” has baggage as a rhetorical term). The Commission should look for low-hanging fruit, particularly with existing government uses, to achieve win-win scenarios when possible.
“Why didn’t Recommendation 5.1 of the National Broadband Plan work?” the group asked. Its answer: Legacy technology on inflexible licensed bands were not always retired rationally. It concluded that three factors seemed to unintentionally encourage this inefficient result:
There are some “informative examples” of these dynamics in action, including the transition envisioned for white spaces in broadcast spectrum, 3.5 GHz, spectrum frontiers, the GPS/Ligado case, LTE-U, 5350-5470, channel sharing via, and broadcast TV.
The group suggested that easing inefficient assignments to obsolete uses may well come through multi-tiered uses. Rather than thinking in terms of “exclusive versus shared use” for these bands, the transition to a new flexible license can be encouraged by shifting to multi-tiered and/or prioritized access. Examples include plans for FirstNet and 3.5 GHz. Non-flexible legacy licenses for specific and obsolete technologies (criteria for specifying these technologies need to be determined) should be considered “transitional bands.” Examples might include AM radio, some radar and 1.7 GHz (DoJ/IRS analog video surveillance to digital). The Commission should actively look for private-public partnership opportunities (technical, political and/or economic) to retire obsolete technologies and make a transition for these bands to the flexible use licensed/unlicensed paradigm. In the interim—which could possibly last a very long time given the particular technologies and applications being used—transitional bands would follow a tiered model to minimize wasted opportunities.
Recommendation No. 3: Rethink universal access.
National leaders generally agree on the aspirational goal of 100 percent broadband access and adoption across the United States. That challenge is beyond the scope of spectrum rules to advance, but the goal of universal inclusion is embedded in the law and the National Broadband Plan. This recommendation is prompted by the rise of new data and new technologies, which require re-scoping the assumptions of the universal access program and its implementation.
The group recommended that technology for serving the remaining unserved areas should remain neutral; fiber is not a realistic requirement for these areas. Fixed wireless, cellular and satellite technologies are now all viable delivery vehicles for broadband, even at speeds currently defined by the FCC. For these reasons, existing programs (such as the Connect America Fund, The Rural Utility Service loan program, Lifeline, etc.) must be re-scoped based on technology improvements and new types of digital divides that have arisen. (Satellite carriers may be impacted by the Frontiers Notice of Proposed Rulemaking.)
The presentation of Working Group B provoked a discussion about whether Recommendation No. 1 could truly work. What would be the standard or policy mechanism for determining the “harm claims threshold” of spectrum interference?
Michael Calabrese of New America theoretically agreed with the recommendation, but complained that it “assumes the problem away” because it does not explain the mechanism by which the FCC is going to retroactively assign a harm claims threshold to each service. For GPS systems, for example, this standard would make existing uses non-tenable, he said. Downes replied that there would be no retroactive enforcement or arbitration for existing uses; the technology—in this case, GPS receivers that are not sufficiently sensitive to their own assigned bands—would simply be retired as soon as possible, perhaps with financial assistance from new entrants in adjacent bands. One participant, citing the GPS/Ligado case, suggested that a dispute could be arbitrated by experts that both parties agree upon; this could ease the transition to a new spectrum-use model.
Richard Bennett of High Tech Forum argued that “it’s not that difficult to define the ‘harm’ standard. For GPS, for example, the harm-point would be if your target receiver is lost. That’s an observable phenomena needed by people navigating by GPS, based on a probabilistic analysis. It is a harm claim that could be backed up with technical analysis.” Bennett cited as another example: the degradation of data rates caused by a service in an adjacent band.
Calabrese objected that the FCC would nonetheless have to define the harm threshold; Bennett disagreed. In any case, the goal would be for disputants to talk to each other to try to resolve any problems, not to look to FCC or courts for adjudication. Given the huge variety of cases that would likely emerge, you might have standards that evolve, similar to case law, suggested Charlie Firestone of the Aspen Institute.
Brian Regan, Associate Bureau Chief of the Wireless Bureau at the FCC, suggested that the framing of the issue might need to be enlarged: “Why not consider flexibility more globally for licensed spectrum, too, regardless of whether the technology is obsolete or not? Policy could provide flexible access to spectrum on a non-interference base for spectrum that has been licensed. And it could allow new uses in licensed spectrum on a non-interference basis. Essentially, the license is a prioritization.”
Calabrese replied, “I thought that was what Recommendation No. 1 is proposing. Even where there are priority access licenses, spectrum would be available for opportunistic uses. As a general rule, that would be the best way to go, so long as no harmful interference to user or licensee.” But Carl Povelites of AT&T Services noted, “This is putting a lot of optimistic faith in SAS technology and opportunistic use. Do they work well? What are the economics behind that?”
Richard Bennett took another view: “Whenever you have legacy systems trying to share spectrum, it will be uncomfortable for someone. We need to think of the spectrum as transition bands where you happen to have one application that will be on the way out. The question is, how do we facilitate that process so that we have flexible use everywhere?” Blair Levin cautioned that we should not necessarily assume rational government budgeting to facilitate this process. He recommended giving consideration to creating a spectrum reallocation fund that could be used for capital expenditure funds.
Working Group B consisted of Larry Downes (coordinator), Richard Bennett, Doug Brake, Andrew Clegg, Alex Hoehn-Saric, Aalok Mehta and Steve Sharkey.
International Issues, Including Satellites
Working Group C presented two recommendations for dealing with international spectrum and satellite issues.
Recommendation No. 1: The US should continue successful implementation of Recommendation 5.16 of the FCC’s Broadband Plan.
This recommendation urges the FCC to promote within the International Telecommunications Union (ITU) innovative and flexible approaches to global spectrum allocation that take into consideration convergence of various radio communication services and enable global development of broadband services. In addition, the group recommended that “the US should show the rest of the world (ROW) how to do this in the context of satellite and terrestrial services.”
The group noted that flexible use policies were adopted by New Zealand in 2007, later by the World Radio Conference (WRC) in 2012 and as Agenda item 1.1 for mobile broadband for International Mobile Telecommunications (IMT). But because there is institutional inertia and incumbents do not necessarily appreciate the value of flexible spectrum service, it is important to implement such practices domestically first, to show that it is feasible, and then try to work through the time-consuming international processes. It may be possible to deal with border issues bilaterally or through regional harmonization, the group suggested.
It is possible no global harmonization of flexible use will be possible, the group conceded. But that may not matter if US markets adopt such approaches. Major markets may scale without global reach. It may be possible for opportunistic spectrum access technologies to overcome the lack of international harmonization, via systems on chips defined by national boundaries. Since technical studies about the feasibility of this approach may be unrealistic, the group suggested that industry standards groups may wish to step in. If the “rest of the world” beyond the US and a few other adopters decline to move ahead, there may be an opportunity for auction revenues for developing countries.
Recommendation No. 2: The US should show the rest of the world (ROW) how to implement innovative approaches to spectrum allocation to ensure maximum flexibility for advanced communications services in the context of satellite and terrestrial services.
The US could lead by example in extremely difficult cases (e.g., “Spectrum Frontiers”) to demonstrate alternatives that “maximize flexibility and rely on market-based solutions to pick highest/best use over long run.” This will require a repeal or refinement of the ORBIT Act auction ban for “international satellite” services.
This recommendation poses a several challenges and opportunities. There are multiple scenarios, such as granting flexible terrestrial and extra-terrestrial rights, and letting markets choose which to use over time. Companies could use existing satellite bands and grant terrestrial rights (e.g., Mobile Satellite Service Ancillary Terrestrial Component rules (MSS ATC)); they could use existing terrestrial rights and grant satellite rights; or they could use new spectrum bands and grant both (e.g., Digital Audio Radio [DARs] terrestrial repeaters).
There are many opportunities but also many technical complexities involved. There are opportunities for earth station protection zones, monitoring, buyout options and short-lived small satellites. While there is a risk of “unjust enrichment,” this could be dealt with via secondary markets. There may be problems with domestic-only satellite footprints as opposed to global coverage, but these might be dealt with via landing rights on a country-by-country basis. It is also possible that higher frequencies could be opened up, which would create an opportunity to create flexible spectrum rights from the core of the earth to the heavens.
Working Group C consisted of Bill Bailey, Michael Calabrese, Blair Levin, Brian Regan, Peter Tenhula and Carl Povelites.
Thus, the world envisioned by the National Broadband Plan in 2010 seems almost quaintly simple compared to today’s spectrum and broadband world, which is experiencing an explosion of new uses, novel technologies and rich opportunities for economic growth, civic and governmental services, and social applications. It is not surprising, therefore, that a recurring if not dominant theme at the Aspen Institute Roundtable on Spectrum Policy is the importance of flexibility. There was a consensus that policy should promote flexible use of spectrum and the sharing of spectrum, and retreat from the historic assignments of fixed spectrum bands for dedicated uses.
This new openness for flexible use reflects many new realities, especially increasing demand for spectrum from new technologies and services, such as drones, autonomous cars and the Internet of Things. Flexible spectrum use also makes sense given the frequently inefficient uses of existing spectrum, the extensive linkages between wired networks and wireless systems, and the emergence of new technologies to facilitate multi-tiered uses.
The challenge ahead is to develop suitable policies to facilitate flexible spectrum use, along with effective implementation of those policies and the technology build-out. To some extent, our ability to more efficiently use spectrum depends upon how efficiently we can expand networks and ‘densify’ network usage. This, in turn, poses many perplexing complications as spectrum policy becomes more implicated in other policies affecting electronic networks. Spectrum policy is not only about signal interference anymore; it increasingly requires coordination with policies involving broadband, net neutrality, privacy, consumer markets, and governmental and national security. This simple reality means that there must be better cross-sector dialogue and coordination as well as more authoritative testing and evaluation of technologies on how spectrum is actually used.
The National Broadband Plan inaugurated a new era of more efficient and productive spectrum usage, which has enabled many new technologies, markets, government services and economic benefits to expand. With the rapid evolution of this space and the change of Administrations, it is a logical time to revisit and refine the policy frameworks for managing the spectrum.