On This Page
Map the fleet and its data handoffs
Bring the flight lead, IT security lead, and person responsible for customer
deliverables into the same review. Use a representative mission to identify the
aircraft, payload, controller or tablet, docking equipment if present, field
network, processing service, storage location, and customer portal. Include
vendor support access and subcontractors.
For each component, record its identifier, software version, responsible person,
connections, and support status. For each data handoff, identify what moves, why
it moves, who can retrieve it, and when it should be deleted. The
NIST Cybersecurity Framework 2.0
provides the broader basis: understand assets and dependencies, assign
responsibilities, and manage access and risk throughout the organization. It
specifies outcomes, rather than a drone configuration recipe.
Keep three flows distinct: flight commands, aircraft status information, and
collected payload data. A command path needs protection against unauthorized
instructions; an image export needs protection against unauthorized disclosure.
They may share infrastructure without sharing the same permissions or
consequences of interruption. The explanation of
secure BVLOS drone communications
develops the link-specific questions.
Consider a proposed roof-inspection workflow: images move from removable media
to a field laptop, then to a processing service and a client folder. Ask whether
the pilot’s phone also caches previews, whether flight logs synchronize
separately, and whether support staff can access the project. These are
questions to resolve for the chosen configuration, not assumptions that every
platform behaves this way.
Understand what each security control protects
Authentication checks that a message or user comes from a trusted source.
Authorization determines what that source may do. Encryption protects
data from being read without the necessary key. Keeping those functions distinct
prevents an encrypted connection from being mistaken for proof that every
connected user should have flight-control privileges.
MAVLink provides a concrete example. Its
message-signing specification
describes how a receiving system verifies signed messages and rejects invalid or
stale packets. It also permits implementation-specific handling of unsigned
messages. Buyers should therefore ask which interfaces enforce signing and which
exceptions remain enabled, rather than treating protocol support as proof of
enforcement.
In the
ArduPilot implementation,
signing does not encrypt telemetry. The documentation also describes USB access
as unaffected by signing and warns that MAVLink 1 serial communication is not
protected by this feature. Those limits make physical controller access and
interface configuration relevant even when command signing is enabled. They do
not describe every commercial drone.
For a broader device review,
NISTIR 8259A
identifies six useful capability areas: device identification, configuration,
data protection, logical access to interfaces, software updates, and
cybersecurity state awareness. Translate them into demonstrations on the
proposed aircraft and controller. Can an administrator identify the exact unit,
restrict configuration changes, protect stored data, disable unnecessary
interfaces, authenticate an update, and retrieve useful security status?
A missing onboard capability may require protection elsewhere in the system, but
that substitute needs its own boundary. For example, protecting a managed laptop
does not establish protection for images left on a removable card. Require the
supplier to identify the storage locations and interfaces its security claims
actually cover.
Build security into the mission workflow
Security tasks should accompany the points where people already prepare
equipment, release a mission, transfer data, and return assets. CISA’s
Secure Your Drone guidance
covers protection before, during, and after flight, including account security,
software provenance, and data handling.
The following proposed handover table applies that lifecycle approach and NIST’s
device and organizational guidance to a commercial assignment. The records are
recommended deliverables, not measured results or universal certification
requirements.
Scroll horizontally to compare all columns.
Before field use, replace default passwords and enable additional account
authentication where supported. Obtain software through authenticated vendor
channels and understand an update before installation, as CISA recommends. Put
updates into the fleet’s maintenance process so the aircraft, controller,
payload, and required application are checked together. If an urgent
vulnerability makes the current configuration unacceptable, hold the affected
equipment until an appropriate change or other mitigation is verified.
After a mission, verify the intended recipient can use the delivered files
before routine deletion of working copies. Apply the agreed retention policy to
the aircraft, controller, removable media, processing workspace, and shared
links. Preserve material needed for an active incident investigation. Offline
collection changes when data moves; examine the later upload and support
workflow before treating it as suitable for sensitive work.
Measure the results you can actually verify
Use repeatable checks to show whether the process works. The measures below are
editorial proposals for a fleet pilot program. Set targets according to the
mission’s consequences and customer commitments, and record the configuration
and observation date with every result.
- Configuration coverage: units with a current, reviewed
aircraft/controller/software record divided by all units eligible for
dispatch. Keep quarantined equipment visible in a separate count.
- Access removal: elapsed time from an approved removal request until a test
confirms that the former user’s relevant access has ended. Check existing
sessions, application integrations, and any supported offline access as well
as a new login. Record any access that could not be revoked remotely.
- Patch exposure: count affected units and the age of unresolved, applicable
vulnerabilities. Distinguish “update available,” “evaluated,” and “installed
and checked.”
- Data closure: completed jobs with confirmed delivery and completed
retention actions divided by jobs due for closure in the reporting period.
- Recovery exercise: elapsed time to restore the selected service or device,
plus a record of which mission functions and data were successfully recovered.
These measures make ownership gaps visible. A fleet can show a high
configuration-coverage percentage while still carrying one overdue vulnerability
on its most consequential mission. Report exceptions beside percentages, with
the responsible person and next action. A low incident count also needs context:
limited monitoring may leave events unseen. Avoid a single security score that
conceals those differences.
Define the demonstration conditions before evaluating a supplier. For access
removal, use an authorized test account and harmless sample data. For recovery,
choose the service whose loss would actually stop the proposed workflow. A
cloud-dashboard restore does not establish that an unavailable controller can be
replaced and returned to flight service.
Prepare for incidents and understand evidence limits
Write down who handles flight safety, who leads the technical investigation, and
who communicates with the customer. A suspicious account login, unexplained
configuration change, lost controller, and missing data export need different
first actions. Exercise those handoffs before relying on them in the field.
NIST SP 800-61 Revision 3
treats incident response as part of continuing cybersecurity risk management. It
emphasizes preserving investigation records, protecting evidence, and verifying
restored assets before normal operation resumes. For a fleet, apply those
principles to flight logs, controller state, cloud activity, affected data, and
the sequence of actions taken.
If an aircraft is airborne, coordinate containment with the flight lead and the
platform’s documented contingency procedure. An unplanned network disconnection
can change the flight situation. Review the aircraft-specific behavior discussed
in UAS lost-link procedures before
designing an incident response that interrupts communications. Conduct
disruptive security evaluations only in an authorized, controlled test
environment.
Keep conclusions proportional to what was evaluated. Protocol documentation
establishes intended behavior. A configuration review establishes selected
settings. A witnessed demonstration establishes what happened under its recorded
conditions. None alone establishes resistance to every attack or continued
protection after later software and integration changes.
Likewise, an assessment of a vendor’s cloud service needs a stated scope: it may
not include the controller, aircraft storage, radio link, or subcontractor
workflow. Ask for the tested versions, interfaces, exclusions, unresolved
findings, and remediation status. Do not infer incident probability or a
universal security ranking from the presence of a report.
Ask these questions before buying or expanding
Ongoing support belongs in the purchase decision.
NISTIR 8259B
complements device capabilities with manufacturer documentation, ways to receive
questions and vulnerability reports, security communications, and education. Its
examples include support terms, update notices, and end-of-support information.
Ask the supplier to answer these questions for the exact proposed configuration:
- Which aircraft, controllers, applications, cloud services, and interfaces
does each security claim cover?
- Can pilot, administrator, maintainer, and customer access be separated, and
what happens to existing sessions when access is revoked?
- Which functions require internet access, which data leaves the field
equipment, and what changes in an offline mode?
- Who can authorize updates, how is authenticity checked, and what supported
recovery path exists if an update fails? Is rollback supported, restricted,
or unavailable?
- Which logs can the fleet export, how long are they available, and what
assistance is available during an incident?
- When does security support end, how are affected customers notified, and how
are data and access removed at retirement?
Put the answers, demonstrations, responsibilities, and continuing support
obligations into the agreed delivery scope. Where a supplier cannot demonstrate
a requirement essential to your mission, resolve that gap before purchase or
wider dispatch. Start with one representative mission and carry its documented
controls through equipment changes and customer delivery. That gives the fleet a
process it can maintain as it grows.
Source notes
- NIST Cybersecurity Framework 2.0:
organizational risk-management outcomes, assets, responsibilities, and access
management; published February 2024.
- MAVLink message signing:
primary protocol documentation for authentication, stale-packet handling, and
unsigned-message exceptions.
- ArduPilot MAVLink2 signing:
implementation documentation distinguishing command signing from telemetry
encryption and describing interface limits.
- NISTIR 8259A:
May 2020 technical capability baseline for connected devices; a starting point
for requirements, not a drone certification.
- CISA, Secure Your Drone: Privacy and Data Protection Guidance:
January 2023 lifecycle guidance, hosted by the FAA Safety Team; used for
account, software, and data-handling recommendations.
- NIST SP 800-61 Revision 3:
April 2025 incident-response guidance, including evidence preservation and
verified recovery.
- NISTIR 8259B:
August 2021 manufacturer support baseline, including documentation,
vulnerability communications, and support lifecycle.
- Featured photograph by Nenad Stojković / Shixart1985:
original photograph and attribution record on Wikimedia Commons; used under
CC BY 2.0. Wikimedia’s
reduced-resolution version is used without cropping or retouching. The
photograph illustrates controller handling, not a security test or commercial
deployment.
Last checked: September 8, 2026.