Modernization Efforts: Why Upgrading Air Traffic Control Is So Hard

Part four of my series on air traffic control. I’m not an ATC professional; this is what I learned from a lot of reading. Sources are at the bottom, and corrections are welcome.

Anyone who has worked on replacing a legacy system knows the phrase “changing the engine while the plane is flying.” In software, it’s a figure of speech. In air traffic control, it isn’t. The system can’t go down for a weekend while you swap in the new version. There’s traffic in the sky every minute of every day, and every change has to be made underneath it.

That one fact explains a lot about why modernization takes so long.

Why Bother?

#

The case for modernizing is simple. Air travel keeps growing, and much of the equipment handling it is old. In January 2023, a failure in the FAA’s NOTAM system (the system that publishes notices to pilots about closed runways, equipment outages, and other hazards) led the FAA to halt all domestic departures nationwide for the first time since September 11. The FAA traced the problem to contract personnel who had unintentionally deleted files while working on the database.¹ A mistake like that shouldn’t be able to stop an entire country’s air travel, but on old, tightly coupled systems, it can.

Beyond reliability, newer technology offers more direct routes (less fuel, fewer emissions), better tracking over oceans and remote areas, fewer voice-radio misunderstandings, and more capacity at busy airports.

A Cautionary Tale

#

Before NextGen, there was the Advanced Automation System. It was the centerpiece of the FAA’s 1981 modernization plan, meant to replace the computers and displays at the heart of air traffic control. In 1983, the FAA estimated it would cost $2.5 billion and be finished in 1996. By 1994, estimates had climbed as high as $7.6 billion, with completion pushed as late as 2003.² That year, the FAA restructured the program, cancelling major pieces outright. Parts of it were eventually replaced by smaller, more focused projects, including the STARS system that TRACONs use today.

Anyone who has spent time in software has seen smaller versions of this story: a huge, all-at-once replacement project, requirements that keep growing, software that turns out to be harder than anyone planned, and eventually a reset. The lesson most people draw from it (and the one I’d draw) is that incremental change beats the big bang. NextGen, for all its problems, was built more along those lines.

NextGen: What It Actually Delivered

#

NextGen launched in 2003 as a long-term program to move the US from radar and voice to satellites and data. Some of its major pieces:

• ADS-B: aircraft broadcast their GPS position to controllers and other aircraft. Since January 1, 2020, it’s been required for aircraft in most controlled US airspace (Class A, B, and C, and Class E above 10,000 feet). It supplements radar rather than replacing it.³
• Performance-Based Navigation: GPS-based routes and approaches that let aircraft fly more direct paths instead of zig-zagging between ground beacons.
• Data Comm: text-based clearances and instructions sent straight to the cockpit, cutting down on radio congestion and misheard numbers. It was in use at 65 airports as of 2025.³
• Time-Based Flow Management: scheduling arrivals by time instead of spacing them out by distance, which cuts down on holding and last-minute vectoring.

None of that is trivial, and the FAA says NextGen has delivered billions of dollars in benefits.³ It has also been slower and more expensive than promised. In 2017, the Government Accountability Office reported the FAA’s own estimate: $20.6 billion in FAA costs through 2030, plus $15.1 billion for the aviation industry to equip its aircraft.⁴ Airlines have to pay for new avionics before they see much benefit, which makes them reluctant early adopters. And the FAA has to keep the old systems running alongside the new ones during every transition, which means paying for both.

SESAR: Europe’s Version

#

Europe’s modernization program, SESAR (Single European Sky ATM Research), is chasing many of the same goals, with an extra layer of difficulty. The US has one air navigation provider. Europe has dozens, each run by a different country with its own equipment, procedures, and politics. EUROCONTROL, the intergovernmental body that coordinates European air traffic management, has 41 member states.

A few SESAR-era ideas have made real progress. Free route airspace lets aircraft fly direct between entry and exit points instead of following fixed airways. Remote towers have gone from concept to reality: Sweden’s Örnsköldsvik Airport has been controlled from a center in Sundsvall since 2015, and London City Airport has run on a remote tower since 2021.⁵ But harmonizing airspace across national borders has been slower than planners hoped, mostly for political reasons rather than technical ones.

Elsewhere

#

Canada’s NAV CANADA was an early customer of Aireon’s space-based ADS-B, which gives controllers a live picture of aircraft over the North Atlantic, where radar can’t reach. ICAO’s Global Air Navigation Plan gives countries a shared framework for modernizing, so that equipment and procedures stay compatible as aircraft cross borders.

What Slows Everything Down

#

Old and new have to coexist. Every new system has to work with the equipment it’s replacing, sometimes for years. That transition period is where cost and risk pile up.

Funding comes in unpredictable pieces. The FAA’s budget is set by annual appropriations, and long-term projects don’t do well with short-term money. Government shutdowns stop work, and budget fights create uncertainty. In 2013, automatic budget cuts under sequestration forced controller furloughs that caused delays across the country, until Congress passed a fix within days.

Security matters more as systems go digital. Moving from isolated radar and radio systems to networked data systems adds new ways for things to go wrong, from equipment failures to deliberate attacks.

People have to be trained on the new thing while doing the old thing. Controllers learn new systems while still working live traffic, and they’re already stretched thin. Pilots and airlines need training too, and procedures have to be standardized across the industry.

Why This Feels Familiar

#

Reading about all of this reminded me of the big software migrations I’ve watched over the years, just with much higher stakes. The technology is usually the easy part. The hard parts are funding, coordination, transition planning, and people. Air traffic control has all four, at a national scale, with zero tolerance for downtime.

Next up: The Future of Air Traffic Control, where I try to separate what’s actually coming from what’s still a slide in a presentation.

References

#