| Internet-Draft | DRIP Reqs | May 2020 |
| Card, et al. | Expires 19 November 2020 | [Page] |
This document defines the requirements for Drone Remote Identification Protocol (DRIP) Working Group protocols and services to support Unmanned Aircraft System Remote Identification (UAS RID).¶
Objectives include: complementing external technical standards as regulator-accepted means of compliance with UAS RID regulations; facilitating use of existing Internet resources to support UAS RID and to enable enhanced related services; and enabling verification that UAS RID information is trustworthy (to some extent, even in the absence of Internet connectivity at the receiving node).¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 19 November 2020.¶
Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
Many safety and other considerations dictate that UAS be remotely identifiable. Civil Aviation Authorities (CAAs) worldwide are mandating UAS RID. The European Union Aviation Safety Agency (EASA) has published [Delegated] and [Implementing] Regulations. The United States (US) Federal Aviation Administration (FAA) has published a Notice of Proposed Rule Making ([NPRM]). CAAs currently promulgate performance-based regulations that do not specify techniques, but rather cite industry consensus technical standards as acceptable means of compliance.¶
ASTM International, Technical Committee F38 (UAS), Subcommittee F38.02 (Aircraft Operations), Work Item WK65041, developed ASTM F3411-19 [F3411-19] Standard Specification for Remote ID and Tracking. It defines 2 means of UAS RID. Network RID defines a set of information for UAS to make available globally indirectly via the Internet. Broadcast RID defines a set of messages for Unmanned Aircraft (UA) to transmit locally directly one-way over Bluetooth or Wi-Fi. Network RID depends upon Internet connectivity, in several segments, from the UAS to the observer. Broadcast RID should need Internet (or other Wide Area Network) connectivity only for UAS registry information lookup using the directly locally received UAS ID as a key. It is expected that the same information will be provided via Broadcast and Network RID; in the US, the FAA NPRM so specifies.¶
[F3411-19] specifies 3 UAS ID types. Type 1 is a static, manufacturer assigned, hardware serial number per ANSI/CTA-2063-A "Small Unmanned Aerial System Serial Numbers" [CTA2063A]. Type 2 is a CAA assigned (presumably static) ID. Type 3 is a UAS Traffic Management (UTM) system assigned UUID [RFC4122], which can but need not be dynamic. The EU allows only Type 1; the US allows Types 1 and 3, but requires Type 3 IDs (if used) each to be used only once (for a single UAS flight, which in the context of UTM is called an "operation"). [F3411-19] Broadcast RID transmits all information in the clear as plaintext (ASCII or binary), so static IDs enable trivial correlation of patterns of use, unacceptable in many applications, e.g. package delivery routes of competitors.¶
An ID is not an end in itself; it exists to enable lookups and provision of services complementing mere identification.¶
Minimal specified information must be made available to the public; access to other data, e.g. UAS operator Personally Identifiable Information (PII), must be limited to strongly authenticated personnel, properly authorized per policy. The balance between privacy and transparency remains a subject for public debate and regulatory action; DRIP can only offer tools to expand the achievable trade space and enable trade-offs within that space. [F3411-19] specifies only how to get the UAS ID to the observer; how the observer can perform these lookups, and how the registries first can be populated with information, is unspecified.¶
Using UAS RID to facilitate vehicular (V2X) communications and applications such as Detect And Avoid (DAA, which would impose tighter latency bounds than RID itself) is an obvious possibility, explicitly contemplated in the FAA NPRM. However, applications of RID beyond RID itself have been omitted from [F3411-19]; DAA has been explicitly declared out of scope in ASTM working group discussions, based on a distinction between RID as a security standard vs DAA as a safety application. Although dynamic establishment of secure communications between the observer and the UAS pilot seems to have been contemplated by the FAA UAS ID and Tracking Aviation Rulemaking Committee (ARC) in their [Recommendations], it is not addressed in any of the subsequent proposed regulations or technical specifications.¶
The need for near-universal deployment of UAS RID is pressing. This implies the need to support use by observers of already ubiquitous mobile devices (smartphones and tablets). Anticipating likely CAA requirements to support legacy devices, especially in light of [Recommendations], [F3411-19] specifies that any UAS sending Broadcast RID over Bluetooth must do so over Bluetooth 4, regardless of whether it also does so over newer versions; as UAS sender devices and observer receiver devices are unpaired, this implies extremely short "advertisement" (beacon) frames.¶
UA onboard RID devices are severely constrained in Size, Weight and Power (SWaP). Cost is a significant impediment to the necessary near-universal adoption of UAS send and observer receive RID capabilities. To accommodate the most severely constrained cases, all these conspire to motivate system design decisions, especially for the Broadcast RID data link, which complicate the protocol design problem: one-way links; extremely short packets; and Internet-disconnected operation of UA onboard devices. Internet-disconnected operation of observer devices has been deemed by ASTM F38.02 too infrequent to address, but for some users is important and presents further challenges.¶
Given not only packet payload length and bandwidth, but also processing and storage within the SWaP constraints of very small (e.g. consumer toy) UA, heavyweight cryptographic security protocols are infeasible, yet trustworthiness of UAS RID information is essential. Under [F3411-19], even the most basic datum, the UAS ID string (typically number) itself can be merely an unsubstantiated claim. Observer devices being ubiquitous, thus popular targets for malware or other compromise, cannot be generally trusted (although the user of each device is compelled to trust that device, to some extent); a "fair witness" functionality (inspired by [Stranger]) may be desirable.¶
DRIP's goal is to make RID immediately actionable, in both Internet and local-only connected scenarios (especially emergencies), in severely constrained UAS environments, balancing legitimate (e.g. public safety) authorities' Need To Know trustworthy information with UAS operators' privacy. DRIP (originally called Trustworthy Multipurpose Remote Identification, TM-RID) potentially could be applied to verifiably identify other types of registered things reported to be in specified physical locations, but the urgent motivation and clear initial focus is UAS. Existing Internet resources (protocol standards, services, infrastructure, and business models) should be leveraged. A natural Internet architecture for UAS RID conforming to proposed regulations and external technical standards will be described in a companion DRIP Architecture document; this document describes only requirements.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
UA may be fixed wing Short Take-Off and Landing (STOL), rotary wing (e.g. helicopter) Vertical Take-Off and Landing (VTOL), or hybrid. They may be single engine or multi engine. The most common today are multicopters: rotary wing, multi engine. The explosion in UAS was enabled by hobbyist development, for multicopters, of advanced flight stability algorithms, enabling even inexperienced pilots to take off, fly to a location of interest, hover, and return to the take-off location or land at a distance. UAS can be remotely piloted by a human (e.g. with a joystick) or programmed to proceed from Global Positioning System (GPS) waypoint to waypoint in a weak form of autonomy; stronger autonomy is coming. UA are "low observable": they typically have a small radar cross section; they make noise quite noticeable at short range but difficult to detect at distances they can quickly close (500 meters in under 17 seconds at 60 knots); they typically fly at low altitudes (for the small UAS to which RID applies in the US, under 400 feet AGL); they are highly maneuverable so can fly under trees and between buildings.¶
UA can carry payloads including sensors, cyber and kinetic weapons, or can be used themselves as weapons by flying them into targets. They can be flown by clueless, careless or criminal operators. Thus the most basic function of UAS RID is "Identification Friend or Foe" (IFF) to mitigate the significant threat they present. Numerous other applications can be enabled or facilitated by RID: consider the importance of identifiers in many Internet protocols and services.¶
Network RID from the UA itself (rather than from its GCS) and Broadcast RID require one or more wireless data links from the UA, but such communications are challenging due to $SWaP constraints and low altitude flight amidst structures and foliage over terrain. Disambiguation of multiple UA flying in close proximity may be very challenging, even if each is reporting its identity, position and velocity as accurately as it can.¶
Network RID has several variants. The UA may have persistent onboard Internet connectivity, in which case it can consistently source RID information directly over the Internet. The UA may have intermittent onboard Internet connectivity, in which case the GCS must source RID information whenever the UA itself is offline. The UA may not have Internet connectivity of its own, but have instead some other form of communications to another node that can relay RID information to the Internet; this would typically be the GCS (which to perform its function must know where the UA is). The UA may have no means of sourcing RID information, in which case the GCS must source it; this is typical under FAA NPRM Limited RID proposed rules, which require providing the location of the GCS (not that of the UA). In the extreme case, this could be the pilot using a web browser to designate, to an UAS Service Supplier (USS) or other UTM entity, a time-bounded airspace volume in which an operation will be conducted; this may impede disambiguation of ID if multiple UAS operate in the same or overlapping spatio-temporal volumes.¶
In most cases in the near term, if the RID information is fed to the Internet directly by the UA or GCS, the first hop data links will be cellular Long Term Evolution (LTE) or WiFi, but provided the data link can support at least IP and ideally TCP, its type is generally immaterial to the higher layer protocols. An UAS or other ultimate source of Network RID information feeds an USS acting as a Network RID Service Provider (Net-RID SP), which essentially proxies for that and other sources; an observer or other ultimate consumer of Network RID information obtains it from a Network RID Display Provider (Net-RID DP), which aggregates information from multiple Net-RID SPs to offer coverage of an airspace volume of interest. Network RID Service and Display providers are expected to be implemented as servers in well-connected infrastructure, accessible via typical means such as web APIs/browsers.¶
Network RID is the more flexible and less constrained of the defined UAS RID means, but is only partially specified in [F3411-19]. It is presumed that IETF efforts supporting Broadcast RID (see next section) can be easily generalized for Network RID.¶
[F3411-19] specifies 3 Broadcast RID data links: Bluetooth 4.X; Bluetooth 5.X Long Range; and WiFi with Neighbor Awareness Networking (NAN). For compliance with this standard, an UA must broadcast (using advertisement mechanisms where no other option supports broadcast) on at least one of these; if broadcasting on Bluetooth 5.x, it is also required concurrently to do so on 4.x (referred to in [F3411-19] as Bluetooth Legacy).¶
The selection of the Broadcast media was driven by research into what is commonly available on 'ground' units (smartphones and tablets) and what was found as prevalent or 'affordable' in UA. Further, there must be an Application Programming Interface (API) for the observer's receiving application to have access to these messages. As yet only Bluetooth 4.X support is readily available, thus the current focus is on working within the 26 byte limit of the Bluetooth 4.X "Broadcast Frame" transmitted on beacon channels. After nominal overheads, this limits the UAS ID string to a maximum length of 20 bytes, and precludes the same frame carrying position, velocity and other information that should be bound to the UAS ID, much less strong authentication data. This requires segmentation ("paging") of longer messages or message bundles ("Message Pack"), and/or correlation of short messages (anticipated by ASTM to be done on the basis of Bluetooth 4 MAC address, which is weak and unverifiable).¶
DRIP WG will focus on making information obtained via UAS RID immediately usable (for the observer to determine whether the UAS is trusted to fly in the airspace volume where and when observed, to establish communications whereby the observer can inquire of the pilot as to intent and/or direct the pilot to exit from the volume, etc.):¶
Any UA can assert any ID using the [F3411-19] required Basic ID message, which lacks any provisions for verification. The Position/Vector message likewise lacks provisions for verification, and does not contain the ID, so must be correlated somehow with a Basic ID message: the developers of [F3411-19] have suggested using the MAC addresses, but these may be randomized by the operating system stack to avoid the adversarial correlation problems of static identifiers. The [F3411-19] optional Authentication Message specifies framing for authentication data, but does not specify any authentication method, and the maximum length of the specified framing is too short for conventional digital signatures and far too short for conventional certificates. The one-way nature of Broadcast RID precludes challenge-response security protocols (e.g. observers sending nonces to UA, to be returned in signed messages). An observer would be seriously challenged to validate the asserted UAS ID or any other information about the UAS or its operator looked up therefrom.¶
Further, [F3411-19] provides very limited choices for an observer to communicate with the pilot, e.g. to request further information on the UAS operation or exit from an airspace volume in an emergency. The System Message provides the location of the pilot/GCS, so an observer could physically go to the asserted GCS location to look for the remote pilot. An observer with Internet connectivity could look up operator PII in a registry, then call a phone number in hopes someone who can immediately influence the UAS operation will answer promptly during that operation.¶
Thus complementing [F3411-19] with protocols enabling strong authentication, preserving operator privacy while enabling immediate use of information by authorized parties, is critical to achieve widespread adoption of a RID system supporting safe and secure operation of UAS.¶
It is highly desirable that Broadcast RID receivers be able to stamp messages with accurate date/time received and receiver location, then relay them to a network service (e.g. SDSP or distributed ledger). This supports 3 objectives: mark up a RID message with where and when it was actually received (which may agree or disagree with the self-report in the set of messages); defend against reply attacks; and support optional SDSP services such as multilateration (to complement UAS position self-reports with independent measurements).¶
A DRIP UAS ID MUST NOT facilitate adversarial correlation of UAS operational patterns; this may be accomplished e.g. by limiting each identifier to a single use, but if so, the UAS ID MUST support defined scalable timely registration methods.¶
Mechanisms standardized in DRIP WG MUST be capable of proving ownership of a claimed UAS ID, and SHOULD be capable of doing so immediately on an observer device lacking Internet connectivity at the time of observation.¶
Mechanisms standardized in DRIP WG MUST be capable of verifying that messages claiming to have been sent from a UAS with a given UAS ID indeed came from the claimed sender.¶
Whether a UAS ID is generated by the operator, GCS, UA, USS or registry, or some collaboration thereamong, is unspecified; however, there must be agreement on the UAS ID among these entities.¶
As satisfying these requirements may require that authorized actors have e.g. Internet connectivity to a Remote ID USS to enable decryption, and such connectivity cannot be assured, DRIP SHOULD provide automatic fallback to plaintext transmission of safety-critical information when necessary.¶
It is likely that an IPv6 prefix or other namespace will be needed; this will be specified in other documents.¶
DRIP is all about safety and security, so content pertaining to such is not limited to this section. DRIP information must be divided into 2 classes: that which, to achieve the purpose, must be published openly in clear plaintext, for the benefit of any observer; and that which must be protected (e.g. PII of pilots) but made available to properly authorized parties (e.g. public safety personnel who urgently need to contact pilots in emergencies). Details of the protection mechanisms will be provided in other documents. Classifying the information will be addressed primarily in external standards; herein it will be regarded as a matter for CAA, registry and operator policies, for which enforcement mechanisms will be defined within the scope of DRIP WG and offered. Mitigation of adversarial correlation will also be addressed.¶
The work of the FAA's UAS Identification and Tracking (UAS ID) Aviation Rulemaking Committee (ARC) is the foundation of later ASTM [F3411-19] and IETF DRIP WG efforts. The work of ASTM F38.02 in balancing the interests of diverse stakeholders is essential to the necessary rapid and widespread deployment of UAS RID.¶