Safeguard
Supply Chain Security•

How Airplanes Get Hacked and How Aviation Is Secured

Aviation cybersecurity explained: how airplanes and airlines get attacked, from GPS spoofing and ADS-B to EFBs and IT outages, and how flights stay safe.

Safeguard Security Team
63 min read

Nobody has publicly documented a cyberattack that took control of a commercial airliner in flight. What aviation does face is real, and mostly lower key. Satellite navigation signals near certain regions are jammed or spoofed on more than a thousand flights a day. Researchers have shown that broadcast radio systems such as ADS-B and ACARS carry no authentication. Crew apps and cabin systems have had real flaws. And ransomware and bad software updates at airline and airport IT providers have grounded thousands of flights. Flights stay safe because aircraft are built in separated layers, every critical function has a backup, and pilots are trained to cross check and fly through a failed system.

That is the short answer. This guide is the long one. It is for airline and airport security teams, avionics and cabin system engineers, maintenance organisations, aviation software suppliers and curious travellers. By the end you will know how an airliner's networks are separated, what runs on its data buses, how each documented class of attack works and why it has or has not mattered, how the industry detects and contains it, what the airworthiness security standards (DO-326A, ED-202A and DO-355A) require, and where the software supply chain fits.

It is written from the attacker's point of view for defenders. For each attack class you will find how it works conceptually, what an attacker would need, real documented cases, the impact, how to detect it and how to defend against it. You will not find exploit code, tool command lines, radio settings or step-by-step instructions. Where research is named, it is named as context for authorised testing and coordinated disclosure.

A note on the law. Testing aircraft, avionics, airline systems or radio navigation and surveillance signals that you do not own, or have no written permission to test, is illegal in most countries. Transmitting on aviation frequencies without authorisation is illegal almost everywhere and can endanger lives. The researchers cited here worked in laboratories, on simulators, on retired aircraft, on their own equipment or through coordinated disclosure.

A word on tone. Aviation is one of the safest ways to travel, and this guide is not meant to make anyone nervous about boarding a flight. The same industry that publishes these incidents is the one that designs aircraft so that no single failure, accidental or deliberate, decides the outcome of a flight. Most of what follows is about how that design holds up and where the weak points actually sit.

This post is part of a device security series. The companion guides on how Wi-Fi gets hacked and how to secure it and how phones get hacked cover the cabin Wi-Fi and the devices passengers bring on board. The guide on how laptops get hacked covers the maintenance and crew laptops that touch aircraft, and the guide on how cars and EV chargers get hacked shows how another safety critical industry separates its networks and signs its updates. Our posts on fabless chip designers and the software supply chain and how a laptop is built explain where the chips and SDKs inside avionics and cabin units come from.

Contents

  1. How does a modern airliner work, and what software does it run?
  2. How are an airliner's networks separated?
  3. What are ARINC 429 and AFDX, and are avionics data buses secure?
  4. What is the attack surface of commercial aviation?
  5. What does an attacker need to target aviation?
  6. Which documented cases shaped aviation security?
  7. Can ADS-B be spoofed, and why are flights still safe?
  8. How do GPS jamming and spoofing affect civil flights?
  9. Is ACARS secure?
  10. Can hackers reach flight systems through in-flight entertainment or Wi-Fi?
  11. How are electronic flight bags attacked?
  12. How is field loadable software protected?
  13. What can go wrong in the parts and software supply chain?
  14. Which airline and airport IT incidents matter most?
  15. How should you weigh these risks?
  16. How are attacks on aviation detected?
  17. What do DO-326A, ED-202A and DO-355A require?
  18. Why is aviation security a software supply chain problem?
  19. How should travellers and crews protect themselves?
  20. How should airlines and airports harden their systems?
  21. How should aircraft makers and avionics suppliers build secure products?
  22. How Safeguard helps aviation software teams
  23. The takeaway
  24. Frequently asked questions
  25. Sources

How does a modern airliner work, and what software does it run?

A modern airliner is a set of dozens of certified computers, each a line replaceable unit (LRU), joined by avionics data buses and grouped into network domains. Flight management, autopilot, navigation and engine control sit in the most protected domain. Crew, maintenance and passenger systems sit in separate domains, and a large ground IT estate keeps the aircraft flying.

Start with the flight deck. The flight management system (FMS) holds the route, the aircraft's performance model and a navigation database of waypoints, airways and procedures. The autopilot and flight control computers turn that route into control surface movements, and on fly by wire aircraft they sit between the pilot's inputs and the hydraulics. Position comes from several sources at once: inertial reference systems (IRS), which use gyroscopes and accelerometers and need no outside signal, GNSS receivers (global navigation satellite systems, such as GPS), and ground radio aids such as VOR and DME. Air data computers supply airspeed and altitude. The transponder answers radar interrogations and, with ADS-B Out, broadcasts the aircraft's position. TCAS, the traffic collision avoidance system, watches nearby transponders and tells pilots how to avoid a conflict. Each engine has its own controller, often called a FADEC (full authority digital engine control).

Messages to and from the airline go through a communications management unit, which routes ACARS datalink traffic over VHF radio, HF radio or satellite. Pilots also carry an electronic flight bag (EFB): a tablet or installed computer with charts, manuals and, on many airlines, the apps that calculate take off and landing performance.

The cabin has its own computers. In-flight entertainment (IFE) servers feed seatback screens and the moving map. A cabin Wi-Fi system and satellite terminal connect passengers to the internet. Crew panels control lighting and announcements.

Diagram of the systems inside a modern airliner: flight management, autopilot and flight control computers, inertial reference, GNSS, air data, transponder with ADS-B and TCAS, communications management unit and engine control joined by ARINC 429 and AFDX buses, plus electronic flight bag, onboard servers and cabin systems, and on the ground airline operations, maintenance, data loaders, passenger service systems, departure control and datalink providers

Figure 1. What runs on an airliner and the ground systems around it.

Every unit runs software, and that software is a controlled part

Most of these units are line replaceable units: boxes a technician can swap at the gate. Their software is treated as a part with its own part number, version and approval. New versions are loaded through a data loader as a loadable software part, which is why the industry calls it field loadable software. Navigation databases on many aircraft are refreshed on a fixed 28 day cycle, and on some older aircraft that still happens with physical media. In 2020 Pen Test Partners showed a retired Boeing 747-400 whose navigation database loader took 3.5 inch floppy disks, refreshed every 28 days from behind a locked panel (The Register). Newer aircraft use network loading, but the principle is the same: software reaches the aircraft as a controlled, traceable part.

The software inside these parts is assured under DO-178C, the long standing guidance for airborne software, while security has its own documents, covered later (PTC).

The ground side is part of the system

An airliner cannot fly a schedule on its own. Airline operations plan flights and assign crews. Maintenance and engineering keep records and hold the library of approved software parts. Passenger service systems run reservations and loyalty programmes. Departure control and airport systems handle check-in, boarding and bags. Datalink providers carry ACARS messages. Most of the incidents that have actually disrupted flights in recent years happened here, in ordinary IT, rather than on the aircraft.

How are an airliner's networks separated?

Airliner networks are divided into domains by trust. The aircraft control domain holds flight critical systems, the airline information services domain holds crew and maintenance functions, and passenger domains hold entertainment, Wi-Fi and passengers' own devices. Gateways, filters and one way data paths are meant to let data flow outward without letting commands flow inward.

The model comes from the industry body that writes avionics standards. ARINC 811, an information report from the Airlines Electronic Engineering Committee published in 2005, sets out a framework for aircraft information security and "provides a common language for understanding security needs" (SAE). Its companion, ARINC 664 Part 5, "defines a set of domains within the aircraft," described "in terms of the services they provide, the security they provide to the functions within their boundaries, and the connections required between a domain and other domains" (SAE). Patent filings that cite ARINC 811 describe its domains as aircraft control (split into avionics and system information), airline information services, and passenger (Google Patents). Researchers commonly describe four: "passenger entertainment, passenger-owned devices, airline information services, and finally aircraft control" (The Register). Naming varies between documents. The idea does not.

Diagram of four aircraft network domains ordered by trust: aircraft control, airline information services, passenger information and entertainment, and passenger owned devices, with boundary devices between them and arrows showing data flowing outward from more trusted to less trusted

Figure 2. The aircraft network domains and the intended direction of data.

  • Aircraft control domain. Flight controls, flight management, navigation, engine control, displays, surveillance. This is the domain whose integrity decides whether the aircraft is safe.
  • Airline information services domain. Onboard servers, crew and maintenance applications, EFB connectivity and datalink routing. It supports the operation but does not fly the aircraft.
  • Passenger information and entertainment services domain. Seatback screens, IFE servers, the cabin Wi-Fi portal and the moving map.
  • Passenger owned devices domain. Phones and laptops brought on board. These are assumed hostile.

The boundaries matter more than the boxes. On the Boeing 787, The Register described "essentially three electronic networks": one for non-critical systems such as entertainment, one for crew and maintenance, and one for avionics. Boeing told the publication that restrictions such as "hardware filters that only allow data to flow between networks rather than instructions or commands" stop one network from directing another (The Register). The moving map is the classic example. The passenger domain needs to know the aircraft's position and speed, so that data flows outward from the control domain. Nothing in the passenger domain needs to send commands back.

Separation is a design property, which means it can be weakened by design choices. Researchers have argued that where a physical path links domains, through a shared satellite terminal for example, the separation depends on how well that path is configured. Makers have argued, with equal force, that their designs do not allow commands to cross. Both positions come up again in the IFE section.

What are ARINC 429 and AFDX, and are avionics data buses secure?

ARINC 429 is a one way avionics bus: one transmitter sends 32 bit words to up to 20 receivers over a twisted pair at 12.5 or 100 kbit/s. AFDX is switched Ethernet with fixed bandwidth virtual links over duplicated networks. Both were designed for determinism and fault tolerance, not to authenticate who is sending.

ARINC 429, the "Mark 33 Digital Information Transfer System," is the predominant data bus on higher end commercial aircraft (Wikipedia). Each word is 32 bits: an 8 bit label that says what the data is, a data field, status bits and a parity bit. "A single wire pair is limited to one transmitter and no more than 20 receivers," and data moves at 12.5 or 100 kbit/s. Labels are standardised by equipment type, so any air data computer reports barometric altitude the same way, which is what lets parts from different makers interchange. Parity catches a corrupted word. Nothing identifies or authenticates the sender, because on a point to point wire inside the airframe the sender is whatever is physically connected.

AFDX (Avionics Full-Duplex Switched Ethernet), defined in ARINC 664 Part 7, is the newer approach. It was developed by Airbus for the A380 and is used on aircraft including the A350 and the Boeing 787 (Wikipedia). Instead of one wire per data flow, end systems send Ethernet frames over virtual links, each a one way logical path with a dedicated bandwidth budget called the bandwidth allocation gap. Switches route frames by virtual link ID and police each link's traffic, so the network's maximum load is known in advance. "A redundant pair of networks is used to improve the system integrity."

Comparison diagram of ARINC 429, a one way point to point bus from one transmitter to several receivers, and AFDX, switched Ethernet with end systems connected to two independent switches A and B

Figure 3. ARINC 429 and AFDX side by side.

What this means for security

Neither bus was built to resist a malicious device already on the bus, and neither needs to be, as long as no untrusted device can get there. That is the job of the architecture around them: physical installation inside the airframe, certified software in every connected unit, domain gateways and the policing that AFDX switches apply per virtual link. The determinism also helps defenders. Traffic on a well defined avionics network is predictable, so unexpected traffic stands out.

The practical attack on an avionics bus is therefore not a radio trick. It is getting malicious or faulty software into a unit that legitimately sits on the bus. That puts the weight on how software is built, signed, distributed and loaded, which is the supply chain topic later in this guide.

What is the attack surface of commercial aviation?

Aviation's attack surface has five parts: radio signals the aircraft trusts (GNSS, ADS-B, ACARS, landing aids), cabin and connectivity systems, crew devices such as EFBs, maintenance and supply chain paths for software and parts, and ground IT for airlines and airports. Only the radio and ground IT groups have seen widespread real world disruption.

Map of the attack surface of commercial aviation in five groups: radio signals it trusts, cabin and connectivity, crew devices, maintenance and supply chain, and ground IT, each labelled as real incidents or research

Figure 4. Where aviation can be attacked, labelled by what has been documented.

  • Radio signals the aircraft trusts. GNSS satellite signals, ADS-B position broadcasts, ACARS datalink and instrument landing system (ILS) beams. All were designed for an era when transmitters were expensive and rare. GNSS interference is a daily operational reality. ADS-B, ACARS and ILS weaknesses have been shown by researchers.
  • Cabin and connectivity. IFE servers, seatback units, the cabin Wi-Fi portal and the satellite terminal. Research has found flaws here, and makers have disputed claims that they lead anywhere dangerous.
  • Crew devices. EFB tablets and the apps and data feeds that update them. Research has shown integrity flaws in performance data, since fixed.
  • Maintenance and supply chain. Field loadable software, data loaders and their media, supplier source code and spare parts paperwork. One real fraud case hit the parts chain in 2023.
  • Ground IT. Passenger service systems, check-in and boarding platforms, crew scheduling and airport operations. This is where ransomware and outages have cancelled flights.

What does an attacker need to target aviation?

It depends on the target. Disrupting airline IT needs the same access as any ransomware attack: phished credentials or an exposed service at a vendor. Interfering with radio signals needs transmitters, proximity and a willingness to break serious laws. Reaching flight critical systems would need several unlikely conditions at once, which is why it has not been documented.

TargetWhat an attacker would needDocumented reality
Airline or airport ITStolen credentials, an exposed service or a compromised vendorReal: ransomware and breaches at shared vendors (2021, 2024, 2025)
GNSS receptionRadio transmitters strong enough to drown out satellite signals over a regionReal: widespread interference reported daily in some regions
ADS-B, ACARS, ILSSoftware defined radio, proximity, protocol knowledge, intent to break aviation radio lawResearch only; cross checks catch false data
Cabin systemsPhysical access in the cabin or network access through the cabin Wi-Fi, plus unpatched flawsResearch and disputed claims; no path to flight controls shown
EFB dataInterception on an untrusted network at the right moment, or local access to the deviceResearch; flaws fixed
Field loadable softwareAccess to a supplier's build, the distribution path or the loading processNo documented attack; protected by signing and process
Spare partsForged paperwork accepted by buyersReal: fraud case (2023)

The pattern is consistent with every other device in this series. The attacks that need the least specialist access, against ordinary IT, are the ones that have actually happened. The attacks that would matter most for safety need the most layers to fail at once.

Which documented cases shaped aviation security?

The cases that shaped the field are a mix of research and real incidents: ADS-B weaknesses shown in 2012 and 2013, cabin and satellite research from 2014 to 2019, the ACARS cipher study in 2017, EFB flaws published in 2024, and real disruption from GPS interference, IT outages and ransomware since 2021.

Timeline of aviation security research and incidents from 2012 to 2025, from ADS-B research in 2012 to ransomware on the MUSE check-in platform in 2025

Figure 5. A timeline of aviation security research and incidents.

Two things stand out. First, nothing on the research side of this timeline led to a documented attack on a flight. Second, the real incidents cluster in two places: radio navigation interference, which has grown sharply since 2023, and ground IT, where a single vendor or a single update can touch many airlines at once. The sections below take each class in turn.

Can ADS-B be spoofed, and why are flights still safe?

ADS-B messages are neither encrypted nor authenticated, so researchers have shown that false aircraft can be injected into receivers. Flights stay safe because air traffic control does not rely on ADS-B alone: it is cross checked against radar and multilateration, and collision avoidance and pilot procedures provide further independent layers.

How it works

Automatic Dependent Surveillance Broadcast (ADS-B) lets an aircraft work out its own position, usually from GNSS, and broadcast it periodically along with its identity, altitude and speed. Ground stations and other aircraft receive it. Most airliners transmit on 1090 MHz using an extended version of the Mode S transponder message (Wikipedia). ADS-B is mandatory for most commercial passenger aircraft in many airspaces.

The design choice that matters for security is simple: "Because the content of ADS-B messages is not encrypted, it may be read by anybody," and "the lack of any authentication within the standard makes it mandatory to validate any received data by use of the primary radar" (Wikipedia). That is why hobbyist flight tracking sites exist. It is also why a receiver cannot tell a real aircraft's message from a forged one by looking at the message.

The research

At Black Hat USA 2012, Andrei Costin and Aurélien Francillon of EURECOM presented Ghost in the Air(Traffic). They examined ADS-B and concluded that attacks ranging from passive eavesdropping to jamming, replay and message injection were "both easy and practically feasible, for a moderately sophisticated attacker," testing against software defined radio receivers and commercial ADS-B receivers (EURECOM).

A year later, Matthias Schäfer, Vincent Lenders and Ivan Martinovic measured the problem more carefully in a controlled experiment published at ACNS 2013. Their conclusion: "attacks on ADS-B can be inexpensive and highly successful," and "any safety-critical air traffic decision process should not rely exclusively on the ADS-B system" (ACNS 2013 paper).

What an attacker needs and the impact

An attacker needs a transmitter on the right frequency, knowledge of the message format, proximity to the receivers they want to fool, and a willingness to commit a serious offence. The impact is on the picture people see, not on the aircraft. Phantom aircraft could appear on a receiver, or a real aircraft's broadcast position could be contradicted. That could add workload for controllers or pilots using ADS-B traffic displays.

Why the impact is limited

The second half of the Schäfer conclusion is the defence. Air traffic surveillance is built so that ADS-B is one input among several. Multilateration compares the arrival time of a transmission at several ground stations to work out where it really came from, so a message claiming a position far from its true source can be flagged (Wikipedia). Primary and secondary radar provide independent tracks. Controllers talk to pilots, who confirm their position. A phantom target that does not answer the radio or appear on radar is a data quality problem, not a collision.

Detection and defence

  • Cross check ADS-B against independent surveillance (radar, multilateration) and flag tracks that do not match.
  • Check message plausibility: impossible speeds, positions that jump, identities that appear in two places.
  • Treat ADS-B traffic displays in the cockpit as awareness aids and keep pilots trained to confirm conflicts by other means.
  • For ground station operators, monitor for sudden bursts of new targets and for anomalous signal strength or angle of arrival.

What about landing signals?

The instrument landing system (ILS) guides aircraft down to the runway in poor visibility using two radio beams, one for left and right and one for the glide path. At USENIX Security 2019, Harshad Sathaye and colleagues at Northeastern University showed "the feasibility of spoofing ILS radio signals using commercially-available SDR," which in a flight simulator's autoland caused offset touchdowns of 18 to over 50 metres (USENIX). The work was done on aviation grade receivers in the lab and in a simulator, not on real aircraft. Pilots flying an ILS approach monitor it against other references, and a go around is a routine, trained response to an approach that does not look right. The research is a reason to keep those cross checks, and a reason for the industry to study authenticated landing aids for the long term.

How do GPS jamming and spoofing affect civil flights?

GNSS jamming drowns out satellite signals and spoofing replaces them with false ones. IATA reports that GNSS interference rates rose 175 percent from 2023 to 2024 and GPS spoofing incidents rose 500 percent. OPSGROUP estimated about 1,500 flights a day were spoofed by summer 2024. Crews handle it with backup navigation and procedures.

This is the one radio threat that is not theoretical. It happens on airline flights every day.

How it works

Satellite signals are extremely weak by the time they reach the ground. Jamming is transmitting noise on the same frequencies so receivers lose the signal. Spoofing is transmitting counterfeit satellite signals that are stronger than the real ones, so receivers lock on and compute a wrong position and time. Jamming is easy to recognise: the GPS simply drops out. Spoofing is harder, because the receiver reports something that looks valid.

The numbers

IATA's 2024 Safety Report said reports of GNSS interference, including signal disruptions, jamming and spoofing, "surged between 2023 and 2024. Interference rates increased by 175%, while GPS spoofing incidents spiked by 500%." IATA named Türkiye, Iraq and Egypt as the areas where interference was most prevalent, and noted that "there are several back-up systems in place to support aviation safety even when these systems are affected" (IATA). In June 2025 IATA added that "the number of global positioning system (GPS) signal loss events increased by 220% between 2021 and 2024" according to its flight data exchange, and reported interference increasing "across Eastern Europe and the Middle East in recent years" (IATA).

OPSGROUP, a membership organisation for pilots, dispatchers and operators, tracked spoofing through crew reports. In July 2024 it reported that the number of affected flights had risen "from an average of 200 daily in the period January-March, to around 900 daily for the second quarter of 2024," with up to 1,350 on some days (OPSGROUP). Its WorkGroup final report in September 2024, built with about 950 participants, estimated that "on average 1500 flights per day are now spoofed, versus 300 in Q1/Q2 of 2024" (OPSGROUP).

Bar chart of OPSGROUP estimates of flights spoofed per day, about 300 in the first half of 2024 and about 1,500 by summer 2024, alongside IATA figures: 175 percent rise in GNSS interference rates, 500 percent rise in GPS spoofing incidents, and 220 percent rise in GPS signal loss events from 2021 to 2024

Figure 6. GNSS interference figures from OPSGROUP and IATA. The figures measure different things and are not directly comparable.

What spoofing does to an aircraft

The surprise for many crews in 2023 was how far a false position could travel inside the aircraft. In September 2023 OPSGROUP published crew reports from enroute flights where a spoofed signal shifted the GPS position by about 60 nautical miles and "in most cases the IRS becomes unusable, VOR/DME sensor inputs fail, the aircraft UTC clock fails, and the crew have been forced to request vectors from ATC to navigate" (OPSGROUP). The reason is integration. Modern inertial systems are updated from GPS to correct drift, so a confident false GPS position can be accepted and blended in.

OPSGROUP later listed other effects: sudden changes in the aircraft clock "from a couple of hours to 8 hour and 12 hour changes" (OPSGROUP), and false terrain warnings. "False alerts, sometimes hours after the spoofing event, are now routine," it reported in 2024, and some go arounds were directly caused by spoofing effects (OPSGROUP). Because ADS-B Out reports the aircraft's own GNSS position, a spoofed aircraft also broadcasts a wrong position to everyone else.

Flow diagram showing a false GNSS signal reaching the aircraft's GNSS receivers and spreading to the flight management system, inertial reference system, aircraft clock, terrain warning system and ADS-B Out reports, with crew countermeasures alongside

Figure 7. How a spoofed GPS position spreads, and what crews do about it.

Why flights remain safe

None of this has the drama its numbers suggest, because aviation never relied on GNSS alone. Aircraft carry inertial systems that work without any outside signal, ground based navigation aids still exist, and air traffic control can see aircraft on radar and give headings. IATA's June 2025 plan with its partners centres on "reinforcing the redundancies that are built into the system," including keeping "a minimum operational network of traditional navigation aids" as a GNSS backup (IATA). Crews brief known interference areas, watch for the signs and switch to other sources.

Detection and defence

  • Crews: brief interference areas before departure; watch for position jumps, clock errors, sensor disagree messages and unexpected terrain alerts; where the aircraft type allows, deselect GPS inputs early; navigate with inertial data and ground aids; ask ATC for vectors; report every event.
  • Airlines: publish type specific procedures with the manufacturer's guidance; analyse flight data for interference signatures; train crews on spoofing as distinct from jamming; feed reports into industry sharing.
  • Manufacturers: improve receiver resilience and spoofing detection; make it easier for crews to isolate GNSS from inertial updating; keep alerting systems robust against corrupted position and time.

Is ACARS secure?

ACARS, the datalink airlines use for operational messages, has no security by default, and most traffic is sent in plaintext over radio or satellite. Researchers found in 2017 that some users who wanted privacy relied on simple substitution ciphers that were easy to break. Secure options exist but cost extra.

How it works

The Aircraft Communications Addressing and Reporting System (ACARS) carries short text messages between aircraft and ground. ARINC introduced it in July 1978, originally to report the "out of the gate, off the ground, on the ground, and into the gate" times automatically (Wikipedia). Today it carries flight plans, weather, engine and equipment health data and, in some airspace, air traffic control messages. Messages travel over VHF, HF or satellite and are routed by service providers such as ARINC and SITA.

The research

Matthew Smith, Daniel Moser, Martin Strohmeier, Vincent Lenders and Ivan Martinovic collected about a million ACARS messages over nine months. Their slides summarise the core problem: "ACARS has no security as standard," and while "post-hoc solutions exist," they cost extra, which deters users (Real World Crypto 2018 slides). In their paper Economy Class Crypto: Exploring Weak Cipher Usage in Avionic Communications via ACARS, published at Financial Cryptography 2017, they showed that some operators who wanted confidentiality, mostly business aircraft, used monoalphabetic substitution ciphers that could be broken with little effort (University of Oxford).

Impact

The documented impact is privacy: anyone with an inexpensive receiver can read most ACARS traffic, and operators who believed they had confidentiality did not. Forged messages are a more serious theoretical concern, but operational messages are checked by crews and dispatchers, and safety related clearances are confirmed through established procedures.

Detection and defence

  • Treat plaintext ACARS as public. Do not send anything over it you would not want read.
  • Where confidentiality or integrity matters, use a standardised, cryptographically sound protection scheme from your datalink provider and avionics supplier, not an improvised cipher.
  • Have dispatch and crews confirm unexpected operational messages through a second channel.
  • Monitor datalink logs for messages that do not match a flight's state or schedule.

Can hackers reach flight systems through in-flight entertainment or Wi-Fi?

No researcher has publicly shown a path from in-flight entertainment or cabin Wi-Fi to flight controls. Researchers have found real flaws in cabin systems and argued that poor separation could create risk. Makers have firmly rejected that any route to flight critical systems exists, citing isolation and one way data paths.

This is the most argued corner of aviation security, so this section presents both sides of four widely reported cases.

Table of four cabin and connectivity cases from 2015, 2016, 2018 and 2019, with what researchers said and how makers and others responded

Figure 8. Cabin research: what was claimed and how makers responded.

2015: claims about seat electronics

In 2015 a security researcher, Chris Roberts, was reported to have told investigators that he connected to IFE equipment under seats on flights between 2011 and 2014 and reached aircraft systems, including a claim about engine thrust (The Register). Roberts said his statements had been taken out of context. Boeing's response was direct: "IFE (in-flight entertainment) systems on commercial airplanes are isolated from flight and navigation systems. While these systems receive position data and have communication links, the design isolates them from the other systems on airplanes performing critical and essential functions." Engineers and experts quoted by Airline Ratings called the claims "fanciful and inconceivable" (Airline Ratings). No public evidence has shown the claimed control.

2016: Panasonic Avionics IFE

In December 2016 Ruben Santamarta of IOActive published research on Panasonic Avionics IFE firmware. He reported flaws that could let an attacker alter what passengers see on seatback screens, spoof the moving map and affect crew cabin functions, and said that "architectural segmentation between critical avionics and operational systems should keep these flaws isolated," while a physical path between domains could, depending on configuration, create risk (Threatpost). He was candid about limits: "We can't say that if we break into one system in the passenger entertainment domain that we would end up in the avionics domain because that's extremely difficult and in certain cases totally impossible."

Panasonic called the claims "inaccurate and misleading," said the "few minor concerns (in no way linked to the control of an aircraft)" had been remediated after validation testing in 2015 and 2016, and disputed the claims about payment card data (The Register). IOActive said it stood by its research. IOActive had reported the issues to Panasonic in March 2015 and published after giving the company time to patch.

2018: satellite terminals reachable from the internet

At Black Hat USA 2018, Santamarta showed that satellite communication equipment used for in-flight connectivity on some fleets could be reached from the internet, and that IoT malware was "inadvertently, performing brute-force attacks against SATCOM modems located onboard an in-flight aircraft" (Black Hat white paper). IOActive worked with the industry through the Aviation ISAC and "confirmed that no critical flight operation systems were affected" (IOActive). The white paper adds: "We can confirm that the affected airlines are no longer exposing their fleets to the Internet."

2019: Boeing 787 crew software

In 2019 Santamarta reported memory corruption flaws in a crew information and maintenance component of the 787, found in code he discovered on a Boeing research server that had been left reachable from the internet. He suggested a theoretical chain from the passenger network through the crew network toward avionics, while acknowledging, "We have confirmed the vulnerabilities, but not that they are exploitable." Boeing said "IOActive's scenarios cannot affect any critical or essential airplane system and do not describe a way for remote attackers to access important 787 systems like the avionics system," and that it verified this in the lab and on an aircraft (The Register).

What both sides agree on

Read side by side, the cases share a premise. The passenger domain is untrusted by design, and the control that matters is the boundary between it and everything else. Researchers argue that boundaries must be verified on each configuration. Makers argue that theirs are verified and hold. For defenders the lesson is practical: patch cabin systems like any other networked product, keep satellite and ground management interfaces off the public internet, and test the boundary, not just the boxes on either side of it.

Detection and defence

  • Keep cabin systems on a supported patch cycle with their makers, and track their components in an SBOM.
  • Restrict management interfaces on satellite terminals and cabin servers to authenticated, non public networks.
  • Monitor cabin network traffic for scanning and unexpected outbound connections.
  • Verify domain boundaries during design reviews and after every cabin modification.

How are electronic flight bags attacked?

EFB apps calculate take off and landing performance and hold manuals and charts, so their main risk is integrity: if the data feeding a performance calculation were tampered with, the numbers could be wrong. Pen Test Partners found two such flaws in Airbus Navblue apps, both since fixed.

An electronic flight bag replaces paper manuals and charts. Many airlines use EFB apps to calculate take off thrust and speeds from runway, weight and weather data. That makes the integrity of the app and its data feed a safety matter, even though the EFB is not part of the aircraft control domain.

Flysmart+ Manager (published 2024)

Pen Test Partners found that NAVBLUE's Flysmart+ Manager app for iOS had App Transport Security, Apple's mechanism that forces encrypted connections, "intentionally disabled, together with any form of certificate validation." The app downloaded databases including take off performance, weight and balance, the Minimum Equipment List and departure procedures (Pen Test Partners).

What an attacker would need: to be on the same untrusted network, such as a layover hotel's Wi-Fi, at the moment a pilot updated the apps, and to know which airline and apps to target. SecurityWeek noted the monthly update cycle and the habit of airlines using the same hotels as the factors that made that timing plausible (SecurityWeek). The researchers' concern was that modified performance data could lead to a tailstrike or runway excursion. Pen Test Partners reported the issue in June 2022. Airbus confirmed it the following month, communicated a mitigation to customers in May 2023 and, in the researchers' words, "resolved the issue."

Flysmart LPC-NG (2024)

A second Pen Test Partners report covered LPC-NG, a Windows EFB application for performance calculations. Data files used by the app could be modified so that it produced wrong take off and landing figures. "The attack requires local access to the EFB," which significantly limits the risk. The researchers recorded a long disagreement over whether this needed a fix, and said the issue was fixed less than two weeks after a regulator became involved (Pen Test Partners). They also noted that similar issues in a Lufthansa Systems and a Boeing EFB app "were responsibly fixed."

Detection and defence

  • Makers: never disable transport security; pin certificates for data feeds; sign performance and airport databases and verify signatures on the device, not only on the server; protect local data files with integrity checks rooted in a key the user cannot change.
  • Airlines: manage EFB tablets centrally; restrict updates to trusted networks or VPNs; keep EFB apps current; include EFB apps in penetration tests and supplier security reviews.
  • Crews: follow standard cross checks between EFB results and flight management computer figures. Pen Test Partners noted that some changes would be caught this way and some would not, which is why integrity has to be built into the software too.

How is field loadable software protected?

Field loadable software is protected by treating every update as a signed, controlled part: suppliers sign software parts under ARINC 835, secure data loaders built to ARINC 645-1 verify the signature before loading, and airlines control how parts are stored and handled. The weak points are unsigned legacy parts, media handling and the supplier's own build.

How it works

A loadable software part is a packaged software build for a specific unit, with a part number. It travels from the supplier to the airline or maintenance organisation, sits in a software parts library, and is loaded into the unit by a portable data loader (PDL) or an airborne data loader (ADL). Every hop is a chance to verify, or a chance for something to go wrong.

Flow diagram of how a loadable software part reaches an aircraft: supplier build, signing under ARINC 835, distribution, airline or MRO parts library, data loader verifying the signature under ARINC 645-1, and the aircraft LRU, with the SBOM travelling alongside and risks at each stage

Figure 9. How field loadable software reaches an aircraft, and where it needs protecting.

The industry's current answer is digital signatures end to end. The Aerospace Industries Association's Civil Aviation Cybersecurity Subcommittee says that "applying a digital signature provides software authenticity and integrity assurance" and strongly recommends it so that only authentic, untampered parts are loaded onto aircraft LRUs. It points to ARINC 835 for signing, with loaders built to ARINC 645-1 verifying those signatures. It notes that ARINC 827 "was designed as a point to point transfer mechanism, and not the end-to-end security that ARINC 835 provides" (AIA, 2022). The latest revision, ARINC 835-2, published in December 2025, covers securing loadable software parts "independent of the transport mechanism or media type, through the use of digital signatures and Public Key Infrastructure (PKI) technologies" (SAE).

What an attacker would need

To push malicious software into an aircraft unit, an attacker would need to compromise the supplier's build, steal a signing key, swap media in an airline's library and get past verification at loading, or misuse authorised access at the point of loading. Each of those is protected by process as well as technology: configuration control, part number checks, maintenance records and the requirement that only approved parts are installed. No such attack on an airliner has been documented.

Detection and defence

  • Sign every software part and verify signatures on loading. Phase out unsigned legacy parts where the aircraft type allows.
  • Protect signing keys in hardware security modules with strict access control and logging.
  • Control media and parts libraries: inventory, hashes, access logs and separation between approved and quarantined parts.
  • Harden data loaders and the laptops that drive them, and keep them off general purpose networks.
  • Keep an SBOM for every software part so a newly disclosed vulnerability can be traced to specific units and aircraft.

What can go wrong in the parts and software supply chain?

Two documented cases show the shape of the risk. In 2023, forged paperwork for jet engine parts from a London distributor led to aircraft being briefly grounded worldwide. In 2019, 787 crew software code was found on a supplier facing server left open to the internet. Neither involved hacking an aircraft, and both were supply chain failures.

Forged parts paperwork

In 2023 airlines discovered that AOG Technics, a London based parts distributor, had supplied CFM56 engine parts with falsified release certificates. The CFM56 powers many older Airbus A320 and Boeing 737 aircraft. Safran's chief executive said the company could not quantify the scale at first and that about 100 engines had been identified with questionable parts (Fortune). In February 2026 Reuters reported that the director had admitted defrauding customers "by falsifying documents regarding the source and condition of engine parts from January 2019 to December 2023," with losses of about 39.3 million pounds (Carrier Management). The 2023 discovery led to planes being briefly grounded worldwide over safety fears.

This is not a cyberattack. It is included because it shows how aviation's supply chain depends on documents that prove where something came from, and how slowly trust unwinds when those documents turn out to be false. Software parts have the same dependency. A software part whose origin and contents cannot be proven is the digital version of an engine part with forged paperwork.

Leaked supplier code

The 2019 Boeing 787 research described earlier began with work in progress code that had been "accidentally left open to the internet" on a research and development server, according to Boeing engineers who spoke to The Register (The Register). Whatever one concludes about exploitability, exposed source code gives anyone who finds it a head start in looking for vulnerabilities.

Open source inside aviation software

Avionics software is tightly controlled, but EFB apps, cabin systems, ground tools, maintenance software and airline IT are built like any other modern software, from open source components and third party libraries. The AIA's 2024 recommendations on SBOMs say "the primary use case for SBoM is to aid vulnerability management" and note that avionics suppliers "may have hundreds to thousands of differing product" configurations, with organisations encouraged to use standard formats such as SPDX and CycloneDX (AIA).

Which airline and airport IT incidents matter most?

The incidents that have disrupted the most flights were ordinary IT failures with an aviation multiplier: the July 2024 CrowdStrike update that crashed Windows machines, ransomware on a shared check-in platform in September 2025, a breach of a shared passenger service system in 2021, and ransomware at a major airport operator in 2024.

The 2024 CrowdStrike outage: an IT dependency lesson

On 19 July 2024 CrowdStrike distributed a faulty configuration update for its Falcon sensor on Windows, which sent affected machines into a boot loop or recovery mode. Globally 5,078 flights, 4.6 percent of those scheduled that day, were cancelled, and several US airlines issued ground stops (Wikipedia). Most carriers recovered within a day or two. Delta did not. Its crew tracking system could not process "the unprecedented number of changes triggered by the system shutdown," and Delta cancelled more than 7,000 flights over five days. Its chief executive put the cost at $500 million (NPR), and a later filing put it at about $550 million (Wikipedia).

No aircraft was affected and nobody was attacked. The lesson is about dependency. One update from one supplier touched thousands of organisations at once, and the airline that recovered slowest was the one whose own scheduling system could not absorb the backlog. Resilience depends on staged rollouts, recovery plans that have been rehearsed at scale, and knowing which systems the operation cannot run without.

The 2025 MUSE ransomware attack: a shared vendor

On 19 September 2025, a cyberattack hit Collins Aerospace's MUSE (Multi-User System Environment) platform, used by airlines to check in and board passengers and track bags. Brussels, Berlin Brandenburg and London Heathrow were reduced to manual check-in and boarding, and Brussels reported cancellations and delays on the first morning (SecurityWeek). RTX, Collins's parent company, confirmed in a regulatory filing that ransomware was involved and said MUSE "operates on a customer-specific network outside of the company's enterprise network" (Cybersecurity Dive).

The attack did not touch aircraft. It touched the shared service that many airlines at many airports depend on, which is why one incident disrupted several countries at once.

The 2021 SITA breach: shared passenger data

In March 2021 SITA, an aviation IT provider, confirmed an intrusion into its Passenger Service System, which handles transactions "from ticket reservations to boarding." Close to a dozen airlines informed passengers, and the number of travellers affected was over 2.1 million, most of them frequent flyer members. Some airlines that did not use the system themselves were affected because loyalty data is shared across alliances (BleepingComputer).

The 2024 Port of Seattle ransomware attack: the airport operator

On 24 August 2024 a ransomware attack forced the Port of Seattle to isolate critical systems, affecting Seattle-Tacoma International Airport. The Port refused to pay and later notified about 90,000 people that their personal information had been stolen (SecurityWeek).

What these incidents teach

  • Shared vendors are concentration risk. Check-in, passenger service and endpoint security platforms serve many airlines at once.
  • Manual fallback is a security control. Airports that could check in and board by hand kept flying.
  • Recovery capacity matters as much as prevention. Delta's slow recovery was a scheduling system problem as much as an endpoint problem.
  • Data shared across alliances is only as safe as the weakest holder.

How should you weigh these risks?

Weigh aviation risks by attacker effort against operational impact. Ransomware and IT dependency failures need little specialist skill and have cancelled thousands of flights, so they come first. GNSS interference is common but managed by procedure. Radio injection and cabin attacks need specialist capability and are contained by cross checks and separation.

Matrix of aviation threats by attacker effort and impact on operations: ransomware on shared vendors and IT dependency outages in the act first quadrant; tampered field loadable software, EFB data and forged parts in plan for; GNSS interference and passenger data theft in manage with procedure; ADS-B or ACARS injection, ILS spoofing and cabin compromise in monitor

Figure 10. How aviation threats compare. Positions are qualitative judgements based on the cases in this guide.

A fair objection: a matrix like this understates rare, high consequence scenarios, and the reason they sit low is that aviation has spent decades building the layers that keep them there. Moving a threat to "monitor" is a statement about today's defences, not permission to remove them.

How are attacks on aviation detected?

Detection in aviation combines crew observation, cross checks between independent systems, flight data analysis, monitoring of ground networks and vendors, and industry sharing through bodies such as the Aviation ISAC and OPSGROUP. Many attacks show up first as disagreements between sources that should agree.

On the flight deck

Pilots are trained to notice when sources disagree: two GPS positions, an IRS and a radio aid, an EFB figure and an FMS figure. Disagree messages, position jumps, clock changes and unexpected alerts are the first sign of GNSS interference. Reporting matters as much as reacting. The scale of spoofing became clear because crews reported it.

In airline operations

  • Flight data analysis. Flight data monitoring programmes can spot GNSS loss and spoofing signatures across a fleet. IATA's 220 percent figure came from shared flight data.
  • Datalink and EFB monitoring. Logs of ACARS traffic and EFB updates can show messages or data that do not match the schedule or the approved version.
  • Software configuration audits. Comparing what is installed on each aircraft against what was approved catches unauthorised or outdated parts.

On the ground

  • Endpoint and network monitoring across check-in, boarding, crew scheduling and maintenance systems, with particular attention to vendor connections.
  • Monitoring for exposed services: satellite terminals, cabin management interfaces and supplier servers that should not be on the internet.
  • Testing restore and manual fallback procedures, because a recovery plan that has never been run will not hold at scale.

Across the industry

The Aviation ISAC is the industry's information sharing and analysis centre, and its site invites researchers to "Disclose A Vulnerability" (Aviation ISAC). IOActive coordinated its 2018 SATCOM disclosure with the A-ISAC, and Pen Test Partners presented its EFB research at an A-ISAC event. OPSGROUP's crew reports turned scattered spoofing encounters into a shared picture. Shared reporting is one of aviation's strongest detection tools.

What do DO-326A, ED-202A and DO-355A require?

DO-326A and its European twin ED-202A define the airworthiness security process for designing aircraft and systems against intentional unauthorised electronic interaction. DO-356A and ED-203A provide the methods. DO-355A and ED-204A cover security after the aircraft enters service, during operation and maintenance.

These documents are written jointly by RTCA Special Committee 216 and EUROCAE Working Group 72, both titled "Aeronautical Systems Security" (RTCA).

  • DO-326A / ED-202A, Airworthiness Security Process Specification. "Originally published in 2010," with DO-326A published in 2014, its guidance "is intended to augment current guidance for aircraft certification to handle the information security threat to aircraft safety" (RTCA). It addresses what the industry calls intentional unauthorized electronic interaction (IUEI), and it sits alongside, not inside, DO-178C and the system development guidance, because aircraft cybersecurity was treated as its own discipline (PTC). In broad terms it asks applicants to plan security work, define the security scope and perimeter, assess threats and risks, build in protections and show those protections are effective. RTCA also lists a newer revision, DO-326B.
  • DO-356A / ED-203A, Airworthiness Security Methods and Considerations. "A set of methods and guidelines that may be used within the airworthiness security process defined in RTCA DO-326A / EUROCAE ED-202A" (RTCA).
  • DO-355A / ED-204A, Information Security Guidance for Continuing Airworthiness. Guidance "for the operation and maintenance of aircraft and for organizations and personnel involved in these tasks," intended to show "that the effects on the safety of the aircraft of information security threats are confined within acceptable levels" (RTCA; EUROCAE).

RTCA's list also includes DO-391 on aeronautical information system security, DO-392 on information security event management and DO-393 on ground systems for air traffic management (RTCA).

What this means in practice

For a supplier, the standards turn security into certification evidence. You need to know what is inside each part, which threats apply, which protections you built and how you tested them. For an airline, DO-355A turns security into a maintenance discipline: managing software parts, data loaders, EFBs, ground support equipment and the people who use them. Both depend on an accurate inventory of the software involved. Our deeper guides on DO-326A and ED-202A certification and the DO-326A supply chain cover the evidence in more detail.

Why is aviation security a software supply chain problem?

Almost every documented weakness in this guide lives in software someone else wrote: avionics parts from suppliers, cabin and EFB apps from vendors, check-in and passenger platforms from shared providers, and endpoint tools from security companies. Knowing what is inside that software, and fixing it quickly, is the core security task.

Look back at the cases. The EFB flaws were in vendor apps. The cabin research was about supplier firmware. The 787 code came from a supplier component. The 2024 outage came from a security vendor's update. The 2025 ransomware hit a shared vendor platform. Even the parts fraud was a failure of provenance in the supply chain.

The chain

  1. Chip and component makers supply processors, radios and the SDKs and drivers that run on them.
  2. Avionics and cabin suppliers build units from their own code, real time operating systems, open source and third party libraries.
  3. Aircraft makers integrate hundreds of units and certify the aircraft.
  4. Airlines and maintenance organisations load, update and maintain software parts for decades.
  5. Ground IT vendors run check-in, passenger, crew and maintenance systems for many airlines at once.

Aircraft fly for decades, so a software part can stay in service long after its components stop receiving fixes upstream. That is the long tail problem. A vulnerability published today in a library compiled into a cabin server or ground tool ten years ago needs someone who knows it is there.

SBOMs are how the chain becomes visible

A software bill of materials (SBOM) lists every component and version inside a piece of software. The AIA's working group, chaired from GE Aerospace with Boeing as vice chair and Collins Aerospace, Honeywell and Gulfstream among its members, recommends SBOMs for vulnerability management and proposes a phased plan that includes a prototype civil aviation SBOM database in SPDX or CycloneDX format (AIA). Combined with signed software parts, an SBOM gives a part both proof of origin and a list of contents, the two things the forged parts case showed aviation cannot do without.

How should travellers and crews protect themselves?

Travellers should treat cabin Wi-Fi as a public network, keep devices updated and use a VPN for anything sensitive. Crews should update EFBs only over trusted networks, follow cross checks, brief GNSS interference areas and report anything unusual. None of this is about flight safety, which does not depend on passengers.

Six layer diagram of how flights stay safe: certified design and domain separation, redundant systems, crew procedures and cross checks, independent surveillance, secure software supply chain, and resilient ground IT

Figure 11. The layers that keep flights safe when one system is attacked.

For travellers

  • Treat cabin and airport Wi-Fi as public networks. Use HTTPS sites and a reputable VPN for work or banking. The Wi-Fi guide explains why.
  • Keep your phone and laptop updated before you travel.
  • Turn off automatic joining of open Wi-Fi networks and Bluetooth when not in use.
  • Use airline apps and websites directly rather than links in unexpected messages, especially during disruption, when scammers impersonate airlines.
  • Turn on two factor authentication for your airline and loyalty accounts. Loyalty data has been exposed through shared systems.
  • Keep seatback screens for entertainment. Do not plug unknown USB devices into them, and do not enter payment details anywhere you would not online.
  • Do not try to probe aircraft or cabin systems. It is illegal, and it will not make you safer.

For pilots and cabin crew

  • Update EFB apps only on networks your airline approves, or over its VPN.
  • Follow cross checks between EFB performance figures and FMS figures every time.
  • Brief known GNSS interference areas and the type specific spoofing procedure before departure.
  • Report GNSS anomalies, unexpected datalink messages and EFB oddities through your safety reporting system.
  • Keep crew devices locked, updated and free of unapproved apps.
  • Report anyone tampering with cabin equipment.

How should airlines and airports harden their systems?

Airlines and airports should treat shared IT vendors as critical suppliers, rehearse manual fallback and large scale recovery, manage EFBs and data loaders as controlled devices, control software parts with signatures and SBOMs, prepare crews for GNSS interference, and share incidents through industry channels.

IT resilience

  • Map which vendors and systems the operation cannot run without: check-in, boarding, bags, crew scheduling, flight planning, maintenance records.
  • Rehearse manual check-in and boarding, and practise recovery at full scale, not just on a test system.
  • Size crew scheduling and recovery tools for mass disruption, not average days.
  • Require staged rollouts and rollback plans from software vendors that run on critical endpoints.
  • Keep offline, tested backups of critical systems and know the restore order.

Ransomware and access

  • Enforce phishing resistant multi factor authentication for staff and vendor access.
  • Segment networks so a compromise of office IT does not reach operational systems.
  • Monitor vendor remote access and remove what is not needed.
  • Patch internet facing services quickly and scan for exposures, including satellite terminals and cabin management interfaces.

Aircraft facing systems

  • Manage EFB tablets with mobile device management, restrict updates to trusted networks and keep apps current.
  • Control software parts libraries: approved parts only, signature checks, hashes, access logging and media controls.
  • Harden data loaders and the laptops that drive them, and keep them off general networks.
  • Require SBOMs from avionics, cabin, EFB and ground software suppliers and monitor them for new vulnerabilities.
  • Apply DO-355A practices to operation and maintenance.

Operations

  • Publish GNSS interference procedures per aircraft type and train crews on spoofing.
  • Use flight data monitoring to track interference across the fleet.
  • Join the Aviation ISAC or a similar sharing body and report incidents.

How should aircraft makers and avionics suppliers build secure products?

Makers and suppliers should design domain separation and verify it on every configuration, sign every software part, protect build systems and signing keys, ship SBOMs, fix vulnerabilities in the open source they depend on without waiting for upstream, and run coordinated disclosure that works with researchers.

Architecture

  • Follow DO-326A and DO-356A from the start, with a defined security perimeter for each system.
  • Keep data flowing outward from the aircraft control domain through filters that pass data and refuse commands, and verify them on each aircraft configuration.
  • Treat satellite and cabin connectivity as untrusted, and verify that no physical path bypasses the domain boundary.
  • Keep management interfaces off public networks by default.

Software

  • Sign every loadable software part under ARINC 835 and support verification in ARINC 645-1 loaders.
  • Never disable transport security in EFB or companion apps, and sign data feeds such as performance databases.
  • Run software composition analysis, static analysis and secrets scanning on every build.
  • Keep build systems and signing keys isolated, logged and protected in hardware.
  • Never leave source code or work in progress builds on internet facing servers.

Supply chain and support

  • Produce an SBOM for every software part and share it with customers.
  • Require SBOMs from your own suppliers, and score their vulnerability management.
  • Plan for decades of support: keep the ability to rebuild and patch components long after upstream projects move on.
  • Publish a vulnerability disclosure policy, respond to researchers promptly and work through the Aviation ISAC when coordination is complex.

How Safeguard helps aviation software teams

Safeguard covers the software supply chain behind aviation products in one platform: SBOMs for software parts, apps and ground systems, binary and image scanning, vulnerabilities ranked by reachability, SCA, SAST and secrets scanning on code you own, fix pull requests gated by your CI, Gold patched packages, supplier risk and compliance evidence.

Safeguard does not sit on an avionics bus, and it does not replace DO-178C assurance or the DO-326A process. It sits where most long lived exposure in this guide comes from: the software, open source components, apps and cloud services that avionics and cabin suppliers, EFB developers, airlines, airports and their IT vendors build and run. Teams that take this seriously often use one tool for source scanning, another for binary analysis, another for SBOM management, a questionnaire process for suppliers and a GRC platform for audits. Safeguard does that work end to end, with one inventory and one vulnerability model, from the first SBOM to the merged fix and the audit record.

Flow diagram of Safeguard for aviation software: source repositories, binaries and images, extracted firmware and supplier SBOMs feed SBOMs, reachability ranked vulnerabilities, code scanning and policies, leading to fix pull requests, Gold patched packages, supplier risk, compliance evidence and SBOM delivery, deployed as SaaS, private VPC or on premises

Figure 12. How Safeguard fits aviation software teams, from inputs to fixes and evidence.

See inside the binaries and images you ship and receive

In the Safeguard console, Integrations > Upload App Binary accepts compiled software: Windows EXE, MSI and DLL files, Android APKs and iOS IPAs, VM and disk images in QCOW2, VMDK, VHD and OVA formats, and container tarballs. Safeguard unpacks the file, identifies embedded libraries and their versions, maps them to known vulnerabilities and records licences, and the result becomes a project with a full SBOM. For aviation teams that covers EFB apps on both mobile platforms, Windows maintenance and performance tools, Linux based images for cabin servers and ground systems, and the containers behind airline and airport services. Raw firmware images for embedded units may need to be extracted into their component files first, and the usual limits of binary analysis apply: heavy obfuscation, encryption or unusual packing can give incomplete results. Our guide to scanning app binaries with Safeguard and our explainer on binary composition analysis go further.

Rank vulnerabilities by what can actually be reached

A cabin server image or EFB app can contain hundreds of components, and many vulnerabilities in them sit in code your product never calls. Safeguard's reachability analysis ranks the vulnerabilities your code can actually reach above those it cannot, and adds EPSS (the Exploit Prediction Scoring System) and known exploited context. For a team whose every change may need re-verification, that ranking is how you decide what goes into the next software part and what can wait, and it gives a DO-326A risk assessment the evidence it needs.

Scan the code you own

For EFB apps, cabin software, ground tools and airline services that you write yourself, Safeguard runs software composition analysis (SCA) across the full transitive dependency tree, static application security testing (SAST) on your own code, and secrets scanning that catches signing keys, API tokens and hard coded credentials before they ship. A disabled transport security setting, a missing certificate check or a signing key committed to a repository are exactly the kinds of issue this catches early. And because a leaked software signing key undermines the whole ARINC 835 chain, secrets in build repositories are worth catching before release.

SBOMs your customers, airlines and auditors can use

Every project produces an SBOM you can export in CycloneDX or SPDX, among other formats, which matches the formats the AIA recommends. That lets a supplier answer an airline's question, "which of our units contain this library version," in minutes rather than weeks. Safeguard's Third-Party Risk Manager lets you request SBOMs from your own suppliers, score them and monitor them continuously, so the supplier side of the chain is visible too. See SBOMs for firmware and IoT devices and our firmware supply chain security guide for more.

Fix it without waiting for upstream

This is where Safeguard differs most from a scanner. For code in your repositories, Griffin, Safeguard's remediation model, opens a fix pull request that your own CI builds and tests before anyone merges it, and your policies decide what can merge. For open source components where upstream has not shipped a fix, or where the fix sits in a version you cannot adopt without a larger change and a fresh round of verification, Safeguard Gold provides patched, hardened builds of those packages, and Safeguard can fork and patch the vulnerable package so your next build pulls a fixed version with the smallest possible change. For a supplier supporting a product for decades, after upstream projects have moved on, that is the difference between a contained fix in the next software part and a major version upgrade nobody can schedule. Closed source components still have to come from their vendor, but the open source around them no longer has to wait.

Compliance evidence from the same data

Safeguard's compliance suite collects evidence for frameworks such as SOC 2 and ISO 27001 from the same scans, policies and fix history. For teams working under DO-326A and DO-355A, the SBOMs, vulnerability monitoring records, risk decisions and fix trail are the raw material that certification reviews, airline audits and customer questionnaires ask for.

Deploy where your code lives

Avionics source code, signing infrastructure and EFB data are sensitive. Safeguard runs as SaaS, in your own VPC, or fully on premises. It can run in Zero AI mode with no language model in the loop, or with a model you choose.

How this compares

Aviation software teams often combine several tools: a source SCA tool such as Snyk or Black Duck, separate binary analysis, an SBOM repository, a supplier questionnaire process and a GRC platform. Those are capable products, and certification tools for requirements and verification do a job Safeguard does not. Safeguard's case is consolidation and remediation across the software itself: one inventory across source, binaries and images; one vulnerability model with reachability; fix pull requests and patched packages instead of tickets; and supplier risk and compliance evidence from the same data. Safeguard can also work alongside existing tools and take over SBOM delivery and fixing.

The takeaway

The image most people have of airplane hacking is someone in seat 23C taking over the controls from a laptop. It has never been shown, and the people who build aircraft have spent decades making sure it cannot happen: separated domains, data that flows outward and not back, certified software, signed parts, redundant systems and pilots trained to fly through a failure.

What aviation actually faces is less cinematic. Satellite signals jammed or faked on more than a thousand flights a day, handled by crews with inertial navigation and a radio call. Broadcast systems from an earlier era that anyone can read, kept safe by cross checks. Crew and cabin apps with ordinary software flaws, found by researchers and fixed. And ground IT, where one vendor's outage or one ransomware attack can strand passengers across a continent.

The common thread is software that someone else wrote. Travellers can protect their own devices. Airlines can rehearse recovery and know their vendors. And makers and suppliers can protect the whole industry, by knowing what is inside every software part they ship and being able to fix it without waiting for everyone upstream.

Try it yourself

Click through these interactive Safeguard demos to see the steps described in this guide.

Test yourself

Has a cyberattack that took control of a commercial airliner in flight been publicly documented?

No. Researchers have found flaws in cabin systems, crew apps and radio protocols, but flight critical systems sit in a separate protected domain and every critical function has a backup. Real disruptions so far have come from GPS interference and ground IT incidents.

Why are flights still safe even though ADS-B is not encrypted or authenticated?

Air traffic control does not rely on ADS-B alone. It cross checks against radar and multilateration, and collision avoidance and pilot procedures add further independent layers.

Why were avionics buses like ARINC 429 not built to authenticate senders?

ARINC 429 and AFDX were designed for determinism and fault tolerance, not to authenticate who is sending. ARINC 429 is a one way bus with one transmitter and up to 20 receivers.

How much did GNSS interference grow from 2023 to 2024?

IATA reports interference rates rose 175 percent and GPS spoofing incidents rose 500 percent. OPSGROUP estimated about 1,500 flights a day were spoofed by summer 2024, and crews handle it with backup navigation and procedures.

Why did airlines cancel thousands of flights on 19 July 2024?

A faulty CrowdStrike configuration update crashed Windows machines worldwide, including check-in, crew scheduling and operations systems. About 5,078 flights were cancelled that day, and no aircraft systems were affected.

How is field loadable aircraft software protected?

Suppliers sign software parts under ARINC 835, and secure data loaders built to ARINC 645-1 verify the signature before loading. The weak points are unsigned legacy parts, media handling and the supplier's own build.

Frequently asked questions

Can a commercial airplane be hacked in flight?

No cyberattack that took control of a commercial airliner in flight has been publicly documented. Researchers have found flaws in cabin systems, crew apps and radio protocols, but flight critical systems sit in a separate, protected domain, every critical function has a backup, and pilots are trained to cross check and fly through a failure. The real disruptions so far have come from GPS interference and ground IT incidents.

Can someone hack a plane through in-flight Wi-Fi or the seatback screen?

No researcher has publicly shown a path from cabin Wi-Fi or in-flight entertainment to flight controls. Researchers have found real flaws in cabin systems and argued that poor separation could create risk. Boeing and Panasonic have said their designs isolate passenger systems from flight systems. Treat cabin Wi-Fi as a public network for your own data.

What is GPS spoofing, and is it dangerous for airliners?

GPS spoofing is transmitting fake satellite signals so receivers compute a wrong position and time. IATA reported a 500 percent rise in spoofing incidents from 2023 to 2024, and OPSGROUP estimated about 1,500 flights a day were spoofed by summer 2024. It adds crew workload and can corrupt connected systems, but aircraft carry inertial navigation, ground aids and radar backup, and crews follow procedures to switch away from GPS.

Can ADS-B be spoofed?

Yes, in principle. ADS-B messages are neither encrypted nor authenticated, and researchers showed in 2012 and 2013 that false aircraft can be injected into receivers. Air traffic control does not rely on ADS-B alone. It cross checks against radar and multilateration, which compares signal arrival times to find where a transmission really came from, and pilots confirm positions by radio.

Is ACARS encrypted?

Not by default. Most ACARS datalink traffic is sent in plaintext and can be read with inexpensive receivers. A 2017 study found that some operators who wanted privacy used simple substitution ciphers that were easy to break. Cryptographically sound protection options exist from datalink providers and avionics suppliers, and anything sensitive should use them.

What are DO-326A and ED-202A?

DO-326A and ED-202A are the joint RTCA and EUROCAE airworthiness security process specification, first published in 2010 and revised in 2014. They guide aircraft and system makers in assessing and protecting against intentional unauthorised electronic interaction. DO-356A and ED-203A provide the methods, and DO-355A and ED-204A cover security in operation and maintenance after entry into service.

Why did airlines cancel flights during the CrowdStrike outage?

On 19 July 2024 a faulty CrowdStrike configuration update crashed Windows machines worldwide, including systems airlines use for check-in, crew scheduling and operations. About 5,078 flights were cancelled globally that day. Delta's crew tracking system could not cope with the backlog, and it cancelled more than 7,000 flights over five days. No aircraft systems were affected. It was an IT dependency failure.

What is an SBOM, and why does aviation need one?

A software bill of materials lists every component and version inside a piece of software. Aircraft stay in service for decades, and their software parts, cabin systems, EFB apps and ground tools contain open source components that receive new vulnerabilities over time. An SBOM lets suppliers and airlines find which units contain an affected component quickly. The AIA recommends SBOMs in SPDX or CycloneDX for aviation vulnerability management.

Sources

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

We email you a link to confirm, and only then add you to the list. Unsubscribe in one click. Privacy

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.