When “full pelt” fails
The British National Air Traffic Services (NATS) control center in Swanwick went down last Friday. Details here.
Here’s what that looks like:
Adhering to the old style of increasing capacity during peak times, by throwing more hardware at the problem, this major air traffic control system went down when additional workstations were added to the system. This simple strategy left thousands of passengers stranded at various airports. The failure lasted approximately 30-60 minutes, disrupting flights through Sunday.
"The system runs pretty much at full pelt. That means as soon as the slightest thing goes wrong the ripples go out enormously," stated former air traffic controller, Martin Clipp, to the BBC Radio 4.
This situation will easily cost tens to hundreds of millions of dollars, major customer complaints, and political ramifications that have been running hot and heavy almost from the moment the failure occurred.
This most recent mess is not the first. Swanwick has seen its share of computer glitches with notable failures in 2002, 2008 and 2013.
Now that things are getting back to normal, and NATS states that safety was not affected, everyone agrees it’s time to revisit these old IT systems. Advice to NATS would be to seek out a platform architected to scale elastically. To build such a system effectively, the underlying operational DBMS has to support these kinds of requirements.
Fortunately the Royal Air Force has its own air traffic control systems or……..
Well, on second thought, let’s not even think about that.