How a Single Technical Failure in NATS Infrastructure Can Disrupt Flights from Glasgow to London and Dublin
100% NEWS | Analysis
8 September 2026
For passengers looking at departure boards at Heathrow, Gatwick, Glasgow, Edinburgh, Manchester or Birmingham on Tuesday, 8 September, the situation may have appeared both simple and inexplicable: dozens of flights delayed, others disappearing from schedules, aircraft remaining at stands, and disruption emerging almost simultaneously across completely different parts of the country.
The obvious question is this: how can aviation appear to “break down” at the same time from Scotland to southern England? The answer is that modern airports do not operate as independent islands. They are connected by a common digital and air traffic management infrastructure. And today, dozens of airports did not fail independently. The disruption originated much higher up the system.
Britain’s air navigation service provider, NATS, confirmed a technical problem affecting its flight processing system. At 13:46, NATS publicly reported an issue primarily affecting departures. It later said it had identified the affected part of the system, and at 16:40 announced that a fix had been implemented and recovery had begun. UK airspace was not, however, fully closed.
By the evening, the consequences had become far larger than the original technical fault itself: more than 600 flights had reportedly been cancelled, while the number of delayed and cancelled movements at major airports approached one thousand. British Airways, easyJet, Ryanair and other carriers were forced to dismantle flight programmes that had already been carefully constructed for the day. But the most important question remains unanswered: what exactly failed — and why can the failure of one system generate consequences on such a scale?
First of All: Britain’s Airports Did Not “Close”
This is the most important terminological distinction. An airport closure normally means that the runway cannot be used, that there is an emergency on the airfield, that weather conditions prevent safe operations, or that some other local restriction has halted movements. That is not what happened today.
NATS lost part of its ability to process air traffic normally, meaning that available capacity — the amount of traffic the system could safely handle — had to be reduced significantly. Aircraft continued to fly. Some flights continued to depart. British airspace remained open. But the number of aircraft the air traffic management system could safely accept, process and transfer between sectors was constrained.
A useful analogy is a six-lane motorway. If five lanes are suddenly closed, the motorway itself technically remains open. But its practical capacity collapses. In aviation, the problem is considerably more serious because aircraft cannot simply be left waiting in an airborne traffic jam indefinitely. The flow therefore has to be restricted before aircraft leave the ground. This is where terms such as ground stop, flow restriction and ATFM regulation become relevant across the wider aviation sector.
What Exactly Is NATS — And Why Can It Affect Almost the Entire Country?
The name NATS historically derives from National Air Traffic Services. But NATS is not “the Heathrow department”, nor is it part of an airline or a government ministry. It is Britain’s principal provider of air traffic services — in professional terminology, an ANSP, or Air Navigation Service Provider.
Air Navigation Service Provider (ANSP) is a public or private legal entity providing air navigation services, managing air traffic on behalf of a company, region or country.
Its ownership structure is unusual because it operates as a public-private partnership. Under the current ownership model: 49% is owned by the UK Government, 42% is held by the Airline Group, 5% is held by employees through a trust, and 4% is held by a Heathrow-related shareholder structure. This ownership model was again confirmed by the UK Government and Parliament in 2026.
NATS itself contains different legal and operational entities. One of the most important is NATS (En Route) plc, or NERL. NERL is the licensed monopoly provider of the principal en-route air traffic services within controlled UK airspace. Its economic activity is regulated by the Civil Aviation Authority, or CAA. A separate business, NATS Services Limited, provides services directly to airports on a contractual basis.
There is therefore an essential distinction between tower control at a particular airport and en-route control across the wider national airspace system. Today’s problem was associated with systemic flight-processing infrastructure, not with a failed control tower at Glasgow Airport.
Where Is British Airspace Actually Managed?
NATS operates two principal air traffic control centres. The first is the Swanwick Centre in Hampshire, in southern England. The second is the Prestwick Centre in Ayrshire, Scotland.
Swanwick manages one of the most complex areas of air traffic in Europe. It houses the London Area Control Centre and London Terminal Control Centre. London Area Control manages en-route traffic within the London Flight Information Region, while London Terminal Control manages the immense volume of aircraft arriving at and departing from the London airport system. Prestwick manages northern airspace, including Scotland, Northern Ireland, part of northern England and a large part of North Atlantic airspace.
British reporting today has linked the failure to flight-processing infrastructure associated with Swanwick. However, one point requires particular journalistic caution: at the time of publication, NATS has not yet released a full technical root-cause report. What is confirmed is the functional location of the problem: the flight processing system.
What has not yet been established publicly is whether the root cause was software code, corrupted or unusual data, a database problem, network failure, hardware, configuration error, a deployment issue, or another technical chain of events. The Civil Aviation Authority is expecting a full report from NATS before deciding whether additional regulatory action is required. It would therefore be premature to say today that a particular programmer caused it, server X failed, or this was a cyberattack.
What Is a Flight Processing System?
This is perhaps the most important technical term in the entire incident. An aircraft does not simply leave Glasgow for Dublin because a pilot points it in the right direction. Before the flight takes place, there is a digital description of the planned journey — the flight plan. It contains the route, waypoints, intended altitudes, timing, aircraft characteristics and other information required by the aviation system.
Within European airspace, another major organisation also plays a central role: the EUROCONTROL Network Manager. For IFR flights within the European network, flight plans are processed through EUROCONTROL’s centralised IFPS — Integrated Initial Flight Plan Processing System. IFPS checks, validates and distributes those flight plans to the relevant national air navigation providers, including NATS.
Flight plan is a document filed by a pilot or flight dispatcher with the local air navigation service provider prior to departure, indicating the planned route, timing, and operational details of the flight.
The national system then has to determine where the aircraft is coming from, where it will enter UK-controlled airspace, which sectors it will cross, where it will leave each sector, which controller will receive responsibility for it, when each transfer is expected, and what future workload will be created in every part of the system. This is no longer merely an aircraft’s electronic itinerary. It becomes part of a mathematical model of future air traffic. If that model becomes unreliable, it is not possible simply to tell air traffic controllers to work it out visually.
A Short Glossary of the Crisis
| Term | Meaning |
|---|---|
| ATC | Air Traffic Control — the direct control of aircraft movements |
| ATM | Air Traffic Management — the broader system used to manage air traffic |
| ANSP | Air Navigation Service Provider — an organisation providing air navigation services; NATS is the principal UK provider |
| NERL | NATS (En Route) plc — the regulated monopoly element of NATS responsible for en-route traffic |
| FIR | Flight Information Region — a defined geographical area of air traffic service responsibility |
| Flight Plan | The digital description of a planned flight |
| Flight Processing System | The system that processes and transforms flight data for air traffic control |
| IFPS | EUROCONTROL’s centralised system for processing and distributing flight plans |
| ATFM / ATFCM | Air Traffic Flow and Capacity Management — balancing traffic demand against available capacity |
| Slot / CTOT | An allocated time window within which a regulated flight is expected to depart |
| Ground Stop | A temporary restriction holding certain departures on the ground |
| Fail-safe | A design principle under which the system moves into a safe state when reliability becomes uncertain, even at the cost of major capacity reduction |
| Fallback | A backup or degraded operating mode |
| Radar | One source of information about where aircraft are physically located; radar and flight-data processing are not the same thing |
Why Can Aircraft Not Simply Continue Flying “Manually” During a Computer Failure?
At first glance, this seems strange. Controllers remain at their positions. Radio communications may still work. Radar may continue to show aircraft. So why not simply continue? Because the radar picture answers the question: “Where is the aircraft now?” Flight-data processing answers many more: “Which flight is this? Where is it going? What route is it following? Which sector should receive it next? Which aircraft will enter this sector twenty minutes from now?”
In a modern air traffic system, one controller does not follow an aircraft all the way from Glasgow to Spain. Airspace is divided into sectors. Aircraft are continuously handed from one controller to another. Automated flight-data processing is therefore part of the infrastructure that makes this vast chain of coordination possible.
When a serious failure occurs, one of the most fundamental principles of safety-critical engineering comes into play: if the system cannot trust the integrity of the data, it must refuse to process it automatically rather than provide potentially incorrect information to a controller. That is why a technical failure may produce hundreds of cancellations rather than an aviation accident. From a safety perspective, that may mean the protective layer of the system worked exactly as intended.
We Have Already Seen an Almost Textbook Example: 28 August 2023
The 2023 incident is crucial to understanding today’s disruption. That failure also involved the NATS flight-plan processing system. What happened was almost improbable: the system received a completely valid flight plan for a service travelling from Los Angeles to Paris. Within it was an unusual combination of route waypoints. Some waypoints around the world share the same letter codes even though they are thousands of miles apart geographically.
The subsequent CAA investigation found that the NATS system encountered a combination involving two identically coded waypoints and a series of other conditions. The software attempted to determine the segment of the route passing through UK airspace and encountered a logically impossible result. The system correctly decided that such data should not be forwarded to controllers. It entered a protective state.
Then the most revealing part happened: the backup system activated. But it was given the same data. As a result, both the primary and secondary systems encountered the same logical problem. Both effectively stopped the automatic processing of flight plans within around twenty seconds. This is a classic example of a crucial reliability concept: common-mode failure. You can have two computers, but if both are running the same code and receive the same problematic input, the existence of the second computer does not necessarily protect you from the first computer’s software or logic error.
Why a Backup System Is Not an Absolute Guarantee
Most people imagine redundancy like this: Computer A fails; Computer B takes over. Real-world resilience is far more complicated. If the problem is a failed power supply, an independent backup server may help. If one processor fails, a backup can help. If one network path is lost, an alternative route can help.
However, if the problem exists in the same version of software, a shared database, an incorrect common configuration, the same incoming data, or a shared external dependency, then the primary and backup systems may fail together. That is why modern critical infrastructure requires not merely redundancy, but also diversity — genuinely different ways of performing the same safety-critical function.
What Happened After the 2023 Failure?
Automatic flight-plan processing stopped at 08:32. NATS moved to manual processing. But humans can physically process flight plans far more slowly than computers. Operations therefore continued, but capacity fell sharply. The automated system was fully restored by 14:32, with restrictions then progressively removed between 15:24 and 18:03.
In other words, the direct technical failure lasted for several hours, but the consequences lasted considerably longer. According to the CAA, more than 700,000 passengers were affected: roughly 300,000 faced cancellations, around 95,000 experienced delays of more than three hours, and a further approximately 300,000 experienced shorter delays. The total economic cost to airlines and passengers was later estimated at around £75 million to £100 million. That incident clearly demonstrates the difference between technical recovery and operational recovery.
Why Can a System Be Repaired at 16:40 While Aircraft Are Still Delayed Tomorrow?
Aviation has a form of operational inertia. Imagine one Ryanair Boeing is scheduled to operate: Dublin → Glasgow → Dublin → Manchester → Dublin → London → Dublin. If the second sector is delayed by four hours, the rest of that aircraft’s programme effectively ceases to exist in its original form.
The aircraft is only half the problem. There is also the flight crew, cabin crew, duty-time limitations, airport slots, parking stands, ground handlers, refuelling, catering, baggage, connections and maintenance windows. Every element has its own timetable. The aviation network after a large-scale disruption therefore resembles not a road network, but an enormous machine containing thousands of synchronised gears. If the machine is stopped for four hours and then restarted, those gears do not automatically return to their previous position. British Airways and easyJet have already warned about displaced aircraft and crews, while the wider transport industry expects knock-on effects to continue beyond the initial technical recovery.
Why Was Glasgow Affected if the Problem Occurred in Southern England?
This question illustrates the networked nature of aviation particularly well. Glasgow is far from Swanwick. Much of the airspace above Scotland is controlled from Prestwick. But a Glasgow-London flight must enter the heavily congested London FIR. A Glasgow-Dublin service may depend on an aircraft that first has to arrive from Dublin, Manchester, London or another airport. A Glasgow-Amsterdam flight is connected to the wider EUROCONTROL network.
Disruption therefore spreads not simply by geography, but through schedule relationships. This is known as network propagation, or a knock-on effect. That is why Dublin Airport can also experience disruption even though NATS is not the Irish national air navigation provider.
So Who “Caused” Today’s Failure?
The answer requires precision. At the operational level, the failure occurred within infrastructure operated by NATS. NATS is therefore the operator responsible for the affected system. That does not mean that the direct cause, or individual responsibility, has already been established.
A proper root-cause analysis must determine whether there was a software defect, a particular combination of data triggered the problem, database state became corrupted or inconsistent, an interface between EUROCONTROL and NATS malfunctioned, there was a network or hardware failure, a configuration problem occurred, a software deployment introduced an unexpected effect, recovery procedures took too long, or primary and backup systems were exposed to the same failure mode. The CAA has asked NATS for a full report. Until that report is published, anything more specific remains a hypothesis.
What Technical Scenarios Are Possible?
If today’s event is treated not as a fully diagnosed case, but as an example of the broader class of failures that can affect air traffic management systems, several scenarios are technically plausible:
- A software logic failure, where the system receives formally valid data, but an unusual combination of inputs creates an unhandled state.
- A data integrity problem where information becomes corrupted, contradictory or incomplete.
- A database synchronisation failure, in which primary and backup systems cease to maintain a consistent state.
- A network or communications failure between individual ATM components.
- A hardware failure or configuration error.
- A common-mode failure affecting primary and backup systems identically.
- A cybersecurity incident, though there is currently no published technical evidence establishing that today’s incident was caused by hostile cyber activity.
Historical Precedents: 2014, 2023 and 2025
Britain Experienced Another Serious NATS Failure in 2014
On 12 December 2014, a further significant failure affected air traffic management systems. The system operated at reduced capacity for several hours, with full capacity restored by around 18:45. According to the CAA, around 450 aircraft were delayed, average delays were approximately 45 minutes, airlines cancelled around 80 flights and the total accumulated delay reached roughly 15,000 minutes. An independent inquiry followed, producing recommendations which NATS completed by late 2016.
The 2023 Investigation
After the August 2023 crisis, the Civil Aviation Authority established an independent expert review. Its final report was published in November 2024, containing 34 recommendations directed at NATS, airlines, airports, the CAA and government. NATS subsequently introduced changes intended to prevent a recurrence of that specific combination of circumstances. However, fixing one particular fault does not make an immensely complex information system immune from every other possible form of failure.
The Disruption of 30 July 2025
Another serious British incident occurred just over a year ago. That time, the cause was described as a radar-related technical issue at Swanwick. The disruption lasted more than four hours and affected Heathrow, Gatwick, Edinburgh, London City and other airports. Unlike the 2023 flight-plan processing failure, this was a different category of technical problem. NATS transferred operations to a backup system, after which flights resumed, although significant time was required to clear the accumulated backlog.
This history explains why today’s event is politically sensitive. It is no longer being viewed by parts of the industry as a single isolated episode, but as another major disruption involving national critical infrastructure. That is why Ryanair has once again called for the resignation of NATS chief executive Martin Rolfe, while Wizz Air has called for deeper reform.
Why Does This Keep Happening in Britain?
The most important answer is that it does not happen only in Britain, but the consequences are particularly visible here. First, London airspace is among the busiest and most complex in the world. Swanwick manages overlapping traffic flows involving Heathrow, Gatwick, Stansted, Luton, London City, domestic traffic, European routes and transatlantic movements. NATS has historically indicated that Swanwick alone processes roughly 5,500 flights on a typical day, while Prestwick handles around another 3,400.
Second, British airports operate at very high utilisation levels. Heathrow, for example, does not have vast spare runway capacity that can instantly absorb a backlog. Third, Britain is geographically positioned as a major aviation bridge between North America and continental Europe. Finally, domestic UK aviation is tightly integrated with the European network, meaning a local technical failure quickly turns into an international event.
Europe Manages Air Traffic as a Network
This is another reason why NATS cannot be understood in isolation from EUROCONTROL. The EUROCONTROL Network Manager Operations Centre continuously balances demand and capacity across the European aviation network. It receives information from hundreds of airlines, dozens of area control centres, airports and states.
If NATS indicates that it can accept only a fraction of its normal traffic, EUROCONTROL cannot allow every aircraft to arrive at the UK boundary at the same time. Some flights therefore receive ATFM delays while still on the ground in Amsterdam, Paris, Dublin, Madrid or Frankfurt. That is why a failure at Swanwick may appear, from the passenger’s perspective, as an aircraft being prevented from departing in another country entirely.
International Parallels: United States and EUROCONTROL Outages
One of the clearest international examples occurred in the United States on 11 January 2023, when the FAA’s NOTAM (Notice to Air Missions) system failed. NOTAM provides safety-critical operational information to pilots. The FAA temporarily halted domestic departures across the United States. The subsequent investigation identified corruption involving a database file while contractors worked on the synchronisation of primary and backup databases.
NOTAM (Notice to Air Missions) is a notice filed with an aviation authority to alert aircraft pilots of potential hazards along a flight route or at a location that could affect the safety of the flight.
EUROCONTROL has also experienced major disruptions. On 3 April 2018, it suffered a major Network Manager systems outage after an automated script cleared operational flight data from live production databases and affected the contingency site. Similar technical tower and ATM system failures are regularly recorded across Europe, from Warsaw to Amsterdam Schiphol.
Governance, Accountability and System Architecture
Air traffic management must simultaneously be available almost continuously, capable of processing enormous quantities of data, compatible with legacy systems, and designed so that safety always takes priority over performance. Software engineering contains an uncomfortable fundamental reality: the more complex a system becomes, the greater the number of possible states that cannot realistically be tested exhaustively in advance.
This also reflects the unique British governance structure. The Civil Aviation Authority regulates NERL economically and oversees the safety of air navigation services. The Department for Transport sets government policy. The UK Government simultaneously owns 49% of NATS, while 42% is held by the Airline Group. Airlines are both customers and indirect co-owners, which explains why debates rapidly move beyond engineering into questions of governance, investment and accountability.
The Central Paradox and Recovery Outlook
From an engineering perspective, if the system detected that data could no longer be trusted and stopped automatic processing to protect operations, the safety layer may have operated precisely as designed. The overriding question is why the fallback architecture did not preserve a much greater proportion of normal capacity.
Regarding operational recovery, technical fixes are only the first step. Delayed inbound aircraft, displaced crews, and duty-time limits create complex scheduling backlogs that take substantial time to resolve across the network.
Today demonstrated just how dependent modern aviation has become on an invisible digital layer. When that system can no longer guarantee the integrity of its data, capacity is reduced to the level that the infrastructure can still manage safely. The upcoming investigation will determine whether Britain has encountered a rare anomaly or a deeper resilience challenge within its national critical infrastructure.
We think you might also like these articles:Global Education Forum 2026 held in Davos
Startup World Cup Championship 2026 Davos: WINNERS
Arvils Pekuless: Reimagining Education for the 21st Century by VisLatvijas Vidusskola, Latvia