FAA Flight-Plan System Has Long History of Problems

The way the FAA has its flight-plan system set up, there simply is no bandwidth for testing a newer system.

"The FAA, with our lives and livelihood in their hands, should be a LOT more proactive in addressing modernization needs before it becomes a crisis with a scope and complexity that defies resolution," server admin Milne wrote to eWEEK.

"This kind of thing [system crashes] happens when changes are pushed into production without adequate testing. It's ironic that the system being criticized is (even after 20 years) the state of the art for functionality worldwide, because the infrastructure that was once redundant now has both platforms fully tasked—much less having resources for a test system," Milne wrote.

The way to port a legacy application is to build an exact replica on a current platform, forgoing the temptation to implement upgrades of any kind. But the age of the system and other limitations do not allow for adequate testing.

"Anything else is unable to function demonstrably in a parallel operations validation scenario that's necessary to establish sufficient confidence to warrant cut-over," Milne wrote.

"Until the FAA bites the bullet and accepts this restriction, we are subject to similar outages while they get their NextGen architecture stood up and functioning. But considering the cost and the limited disruption up to now, it may be worth accepting the pain for a time. If that's what they want to do, the right thing is to establish a deadline, after which they are obliged to stop and update the legacy application as the price of gaining the time to perfect the new system," Milne wrote.

"My interest is based on being a consumer of the FAA's services—a little over 2 million miles flown so far. I am shocked, no, I'm way beyond shocked at how antiquated the equipment is (where does someone go to get replacement vacuum tubes?). Sure, it works well most of the time (thank God). But we have better technology on golf carts, it seems."