The failure of a link between different systems at Swanwick air traffic control centre in Hampshire delayed flights at major airports including Heathrow, Gatwick and Luton yesterday, and hit some flights today.
National Air Traffic Services, which runs the Swanwick centre, has apologised to travellers. Ian Hall, NATS' director of operational performance, said, "Our systems are incredibly resilient. We take every step to avoid any problems but are always aware that in maintaining and updating highly complex systems, we can experience difficulties." He said there was a backlog of flights as a result of the problems.
When the link failed, NATS issued instructions for some flights not to take off, which reduced the workload on air traffic controllers who reverted to some manual processes to manage aircraft.
The problem is understood to have been traced to a communications sub-system between the IBM-based National Airspace System, sometimes called the National Flight Data Processing System, and Swanwick's separate, more modern area control technology, which was built by NATS and Lockheed Martin. The failure happened at 3.55pm. The full systems were back online by 7pm but some flight cancellations and disruption at airports have continued until today because the sequencing of aircraft take-offs had been disrupted.
There have been serious failures before of links between Swanwick's systems and the National Airspace System [NAS] - but this was the first of its kind. Engineers at NATS are still investigating the exact cause.
The NAS runs on modern IBM hardware but parts of the software date back decades. It is updated infrequently to keep it from becoming unstable. The NAS software was due to have been replaced in 2005 but this has not happened, mainly because NATS is developing a new system with Spanish and German air traffic control specialists.
Links between Swanwick's systems and the older NAS technology are likely to be complex and purpose-built. NATS emphasises that the failure was not of the NAS system itself.
The failure meant that air traffic controllers did not have the information on their flight data strips updated automatically. Although the correct information was in the NAS system, it was not transferring to Swanwick's area control system. So updated flight route information - including waypoint data - had to be input manually directly into Swanwick's area control system.
The flight strips are used particularly when controllers receive flights from other airspace sectors into their own areas. They were able to see the position of aircraft on their screens from the radar "blips". They also had historic text data on their screens of the aircraft and its destination.
But flight strips are used by controllers, in part, to see new or updated instructions which have been given to pilots. So controllers know what aircraft are about to enter their airspace sector and routing information. When the sub-system failed, controllers lost the ability to have recent instructions to pilots updated automatically. When the information had to be handled manually, this reduced the number of flights controllers could manage.
Hall said yesterday: "Safety has not been compromised at any stage and we sincerely apologise to those who have been inconvenienced this evening. We are working closely with airlines to increase capacity this evening to help minimise delays and impact on the rest of today's schedules."
Controllers reported no particular difficulties in reverting to some manual procedures. But airlines reported scores of delayed or cancelled flights and some passengers missed flights because of long queues at check-in desks. British Airways said the further cancellations on Friday affected only short-haul flights. They were the result of a knock-on effect of 35 cancelled domestic and European flights on Thursday, said BA.