The failure of a single workstation at the new £623m air traffic
control centre at Swanwick, Hampshire, caused about 100 other
workstations to crash, leading to long delays at airports and the
cancellation of dozens of flights last Friday.
Although National Air Traffic Services (Nats), which owns the
Swanwick centre, had built costly and elaborate resilience into the
systems, it appears that there was no expectation that a single
workstation could cause so many others to fail.
The problem began at about 6.30am on Friday when Nats tried to boot
up about half of Swanwick's 200 IBM Unix-based workstations that
had been switched off during the night. A single workstation which
had had a routine software update installed blocked the booting up
of the 100 workstations.
This forced Nats to operate at 50% of capacity. To allow
controllers on the remaining 100 workstations to cope, some
aircraft were instructed not to take off until the problem was
circumvented and the systems were back to normal at about 11am.
The disruption was compounded by an unconnected failure of systems
at the main European air traffic control hub, Eurocontrol in
Brussels.
The systems crash at Swanwick and complaints about the lack of
clarity of some display screens raise the question of whether the
Civil Aviation Authority (CAA) should have certified the systems as
safe to go live in January.
Swanwick controls about 5,000 flights a day over England and Wales
from IBM RS/6000 workstations that run the AIX operating system.
The Lockheed Martin software contains more than two million lines
of code.
A spokesman for the CAA said his organisation is in no doubt that
the Swanwick systems are safe. A Nats spokesman said the exact
cause of the workstation's crash and why it affected so many
systems is still being investigated.