Julie Knapp’s comment at a recent FCC TAC meeting have inspired me to sketch out some principles/guidelines/framework for the FCC when looking at protecting services during rulemakings. Julie pointed out that the TAC Spectrum and Receivers Working Group now has an opportunity to synthesize and make actionable our work of the last few years.
Harmful interference arises due to the interaction of two services. It is a function of the characteristics of both the interfering service and the affected service. The characteristics of the affected system, notably its receiver characteristics but also other attributes like transmitter power and location topology, can contribute to harmful interference. Therefore, both the interfering and affected systems share the responsibility to avoid harmful interference. To date, however, regulation has focused on the transmissions of the interfering service and ignored the ability of the affected service to reject interference.
Affected non-safety services are protected against “[i]nterference which … seriously degrades, obstructs, or repeatedly interrupts” their operation (47 CFR 2.1 definition). They are not protected against any interference whatever, just interference that rises to this level. Services should therefore expect and plan for non-harmful interference to their service from current or future adjacent allocations. Affected services should bear in mind that the general definition of interference includes “any performance degradation, misinterpretation, or loss of information which could be extracted in the absence of such unwanted energy”; they should therefore expect and plan for performance degradation, misinterpretation, or loss of information which does not rise to the level of serious degradation, obstruction, or repeated interruption of their service.
To fairly partition the operation of interfering transmitters and affected receivers, interference assessments should by default be done at the receiving antenna input, not at or beyond the receiver input. This default may be over-ridden where entrenched custom or international norms require, such as in protecting high gain satellite earth station receivers.
Protection of Reception
Since receiver characteristics are one determinant of whether a service experiences harmful interference, the Commission should expect that a service’s receivers, and operations generally including transmitter power and deployment, meet industry standards or guidelines. It should base its assessment of the probability and impact (i.e. the risk) of harmful interference on this expectation. Where industry standards or guidelines do not exist, the Commission should base its assessment on the performance of the majority of receivers in the affected service, not the poorest-performing one. (One could be more prescriptive here, e.g. specifying that the FCC should not protect receivers falling, say, more than one standard deviation below the mean on key receiver characteristics, listed below.)
Operators in the affected service are expected to disclose the relevant standards, guidelines or operating (e.g. receiver) characteristics if they require protection. Neither the Commission nor an interfering service can protect what it does not know about. Where receiver and system operating characteristics are disclosed, they should include at least the following (cf. “Table 20. Example of a Generic Receiver Spectrum Standard” in NTIA Report 03-404)
- Receiver characteristics such as spurious response rejection, adjacent channel rejection, semi-adjacent channel rejection, 60 dB selectivity, intermodulation rejection, image frequency rejection, blocking immunity, and minimum and maximum sensitivity (i.e. dynamic range)
- System characteristics such as SINR and margin minima, transmitter deployment, antenna characteristics, and operating bandwidth.
As a guide to the approach the Commission may take in its analysis, one could start with the following general principles and adjust them to reflect the specifics of a particular case:
- An existing service should not expect protection from interference, such as limits on an interferer’s transmit power or resulting interfering field strength, at more than a reasonable frequency offset from the boundaries of the service’s assignment; the boundary in question may be a channel or a band edge, depending on the case. As a rule of thumb, a service should not expect out-of-assignment protection at more than (say) three times the bandwidth of its assignment from the edges of its allocated band (cf. the NTIA Report’s “60 dB Selectivity” requirement).
- Unless the contrary is shown to be required, the Commission should assume that a receiver’s selectivity follows the same (inverted) roll-off pattern as the transmit mask of its desired transmission. (Obvious exceptions include TV transmit and receive masks; the obvious test case is GPS, but radar probably also falls in this category.)
The performance of a receiving system is important in ensuring that harmful interference does not occur. While the Commission is unlikely to specify receiver performance characteristics, it can use interference limits policies, i.e. ways to describe the environment in which a receiver must operate without necessarily specifying receiver performance. For example, it may specify harm claim thresholds, i.e. in-band and out-of-band interfering signal levels that must be exceeded before a system can claim that it is experiencing harmful interference; see FCC TAC (2014) "Interference Limits Policy and Harm Claim Thresholds: An Introduction". Ideally, system characteristics and harm claim thresholds will be agreed by contending parties using multi-stakeholder processes; the Commission could establish default thresholds that stakeholders are free to change and improve by mutual agreement.
To facilitate affected services in bringing enforcement actions against interferers, the Commission may designate conformance to a specific receiver performance standard as a safe harbor for asserting that a harm claim threshold has been exceeded; that is, if a receiver meeting such a standard suffers severe performance degradation, the harm claim threshold will be deemed to have been exceeded without measurements or modeling being required to validate the claim.
Worst case and risk assessment
RF signals, both wanted and unwanted, fluctuate everywhere and all the time due to the physics of electromagnetic propagation. Further, the protection criteria of repeated interruption and serious degradation imply extended and widespread service loss; occasional degradation and interruption at various places and times is unavoidable and does not rise to the level of harmful interference. Protection against harmful interference is therefore a statistical matter. Rules can at most protect services against harmful interference in most places, most of the time; guaranteed, universal protection is unattainable and should not be expected.
Specifically, the Commission should not make allocation decisions on the basis of worst case failure scenarios. It should make risk-informed interference assessments that take into account both the consequence and likelihood of interference hazards; see e.g. FCC TAC (2015) "A Quick Introduction to Risk-Informed Interference Assessment." (Of course, in the final analysis the Commission is required to act in the public interest, which includes but goes beyond technical considerations.) It should not protect against the worst case, but similarly it should not afford protection only against harmful interference hazards based on average values of interference parameters; the offered protection will be somewhere between these two extremes.
Parties petitioning the Commission regarding harmful interference are encouraged to provide quantitative risk assessments. If a party bases its arguments on a single harm scenario, e.g. a worst case, it should provide evidence not just for its impact but also estimate the likelihood of the hazard.
Taking a broad perspective
Wireless systems are vulnerable to many failure modes beyond those attributable to interference from other services. Examples include RF noise from unintentional radiators, incorrect configuration, power failure, excessive link range, self-interference, etc. Interference from other services can be a low probability cause of failure compared to these baseline hazards. The Commission should assess the relative risk of harm due to a new allocation by comparing it to a baseline, i.e. the incidence and severity of pre-existing causes of service degradation, and may require those seeking protection to submit data on baseline outages.
The Commission should take a holistic view of protection against harmful interference: affected systems should be expected to use interference mitigation methods not only at the radio layer in the specific assignment in question, but to use a variety of techniques (where available) including antenna patterns that apply nulls to sources of interference, using alternative radio channels, mitigation higher up in the communication stack such as error correction protocols at the link layer, and repositioning receivers and interfering transmitters both in space and frequency The mere fact that a particular channel may be unavailable due to interference need not amount to service interruption or degradation to the extent of harmful interference if alternatives and/or mitigations are available.
While this came up in the context of the receiver characteristics, and reception is a key part of the new thinking, it’s about more than just receivers; rather, it’s about expectations of protection against harmful interference more generally, and so also encompasses interference limits policy and risk assessment.
I’m focusing for now on adjacent channel/band allocation, not co-channel sharing.
I haven’t tried to connect it with enforcement.
I’m focusing for now on non-safety services, leaving aside for later refinement the inclusion of “radionavigation service or [of] other safety services” that are entitled to protection against their “functioning” being “endanger[ed]” (47 CFR 2.1 definition of harmful interference)