thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $59
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

Blog  /  News

When the Network Grounded a Nation: Inside the Nov 13 FAA Ground Stop

On Friday afternoon, November 13, the FAA stopped every departing flight in the United States. This wasn't weather, and officials say it wasn't an attack. The FAA's preliminary findings point to a few routers on a telecom network most travelers have never heard of. If you work in networking, this is the most important incident of the year to understand. Almost every failure in it is something we teach people to prevent.

A note before we start: the root-cause investigation is still open. The FAA and its network contractor say they're still investigating, and officials have given different accounts of exactly what triggered the failure. Below, we separate what's been confirmed from what's still preliminary, and we label our own analysis as analysis. 3:54p.m. ET: nationwide ground stop issued ~4h 20muntil departures resumed at 8:15 p.m. ET 17 of 22en-route centers (ARTCCs) saw instability 9,159U.S. flights canceled Friday (FlightAware)

The timeline

Click any event for details.

~3:30 p.m.Network instability beginsThe FAA says radar and radio communications began behaving inconsistently at 17 of the 22 Air Route Traffic Control Centers, the facilities that handle high-altitude, en-route traffic. The problem then spread to several TRACONs, which handle approach and departure traffic. Controllers reported screens flickering or going dark and static on voice lines. 3:54 p.m.Nationwide ground stopWith roughly 3,100 to 3,500 aircraft already airborne (estimates vary by source), the FAA halted all departures nationwide "out of an abundance of caution." Flights already in the air kept flying and landed on the radio frequencies and radar feeds that stayed up. EveningProblem sites isolatedAccording to the FAA's initial investigation, the trouble traced to a radar site in Amarillo, Texas, and a radio communications site in El Dorado, Arkansas, where contractor work, including a new router, had been under way. Both sites were blocked off the network, broadcast was disabled on the new routers, and about 60 similar routers were taken offline as a precaution. The network stabilized. 8:15 p.m.Ground stop liftedDepartures resumed, but by then airline crews and aircraft were out of position nationwide. That's why cancellations kept piling up on Saturday and Sunday. Sat–SunThe long tailThousands more flights were canceled over the weekend. Reports put the combined Friday and Saturday cancellations above 14,000. The FAA Administrator expected airlines to recover their full schedules by the end of Sunday.

What actually failed: the FAA's "backbone" network

Air traffic control isn't one computer. It's hundreds of facilities (radar sites, radio sites, towers, TRACONs and en-route centers) that all have to share data in real time. The network that ties them together is the FAA Telecommunications Infrastructure (FTI), operated for the FAA by contractor L3Harris. CBS reports the contract dates back to 2002.

Much of FTI still rides on TDM (time-division multiplexing), the older circuit-based technology that predates modern IP networking. A 2023 safety review commissioned by the FAA warned that the network was obsolete and that an unexpected failure could cut off an individual facility or even a whole regional center. The FAA and L3Harris have been migrating sites from copper to fiber. L3Harris says that conversion is more than 70% complete.

FTI backboneshared transport Radar site (TX) Radio site (AR) Other remote sites En-route center En-route center TRACON Normal dayNov 13: the stormContained by design

Remote sites send radar and radio traffic across a shared backbone to the centers that need it.

How a router change can stop a country

Officials have described the mechanism the same way every time. Routers became overloaded with data and began sending inconsistent information, with more than 1,000 connection attempts every four to five seconds, according to the Transportation Secretary and the FAA. Radar and voice depend on steady, predictable delivery. When the transport underneath them starts flapping, controllers can't trust what they see and hear.

The fact that the fix included disabling broadcast on the new routers is telling. Our analysis, not an official finding: that symptom pattern looks like a classic flood or "storm" failure. A device starts sending traffic that other devices must process. Those devices respond or retry, the retries create more traffic, and control-plane CPUs on older hardware get overwhelmed. Healthy links get dragged down alongside the bad one, because nothing stopped the problem at the edge where it started.

What triggered it? The accounts still differ

Transportation Secretary

A contractor replacing copper with fiber at a remote site, "either through contractor error or a faulty piece of equipment."

DHS Secretary (Friday)

New security patches on the older system appeared to overwhelm its routers.

FAA Q&A

A routine replacement at a remote facility, made worse by the legacy TDM system. Root cause still under investigation with L3Harris.

CBS (preliminary sources)

A network update caused a reboot and crash, limited to an older router family from the early 2000s that has since been identified and isolated.

On cyber, every official account says the same thing: no evidence of a cyberattack and no link to AI tools. A union for FAA technicians has also raised concerns that its technicians weren't included in the upgrade work. Expect that question to come up in congressional review.

Why a ground stop was the right call

It's easy to be angry about the cancellations. But look at it from the controller's seat: when radar returns and voice lines are unreliable, you can't safely add more airplanes to the sky. The FAA kept enough frequencies and radar feeds up to land everyone already flying, along with critical, law-enforcement and military flights. Officials say everyone landed safely. In engineering terms, the system failed safe. The problem is that it failed at all, and that the failure spread so widely.

How this could have been avoided

None of what follows is exotic. It's standard practice in well-run enterprise and carrier networks, and most of it appears on Cisco and CompTIA exam blueprints.

1. Change management and maintenance windows

The work happened in the middle of a weekday afternoon, one of the busiest times in the national airspace. The FAA has said it will re-evaluate daytime maintenance. Every change to a production network should come with a documented plan, a peer review, a defined low-traffic window, success criteria and a tested rollback plan.

2. Stage the rollout (canary first)

A new device configuration should first go into one isolated site, then soak under monitoring before it's repeated anywhere else. The precautionary shutdown of about 60 similar routers suggests the same risk existed at many more sites.

3. Hardened configuration templates

Broadcast and other unnecessary services should be disabled by default on new routers. They should never be enabled by accident. A golden config template plus automated compliance checks (the kind of thing taught in network automation and programmability courses) catches this before a device ever goes live.

4. Contain the blast radius

Storm control, rate limiting and control-plane policing (CoPP) protect a router's CPU from floods. Network segmentation keeps one misbehaving site from dragging down 17 centers. A well-designed network has failure domains: when something breaks, it breaks small.

5. Real redundancy, not shared fate

Backups only help if they don't share the failure. Primary and backup paths that ride the same transport, the same router family or the same configuration will fail together. Senators on the aviation committees are already asking exactly this about primary and backup frequencies.

6. Retire hardware before it retires you

Routers built in the early 2000s are long past vendor support. Lifecycle management (tracking end-of-sale and end-of-support dates and funding replacements on schedule) is less exciting than a $50 billion modernization plan, but it's what keeps a system from reaching this point.

7. Keep the people who know the system in the loop

If the technicians who maintain a network aren't part of the change, nobody in the room knows its quirks. Good change control requires sign-off from the people who will be paged when it breaks.

How tech understanding would have helped

Here's how the concepts in this incident map to what certification students actually learn:

What happenedThe skill that prevents itWhere you learn it
Traffic flood overwhelmed routersBroadcast domains, storm control, CoPP, QoSCCNA, CCNP ENCOR
Bad config reached productionConfig templates, automation, validationNetwork Programmability, ENCOR automation domain
One site took down many centersHierarchical design, segmentation, failure domainsCCNP ENCOR, ENSLD
Midday change, no clean rollbackChange control, documentation, troubleshooting methodNetwork+, CCNA
Was it a cyberattack?Incident response, log analysis, threat triageSecurity+, CySA+, CCNP Security
Legacy TDM to IP and fiber migrationWAN technologies, physical layer, transportNetwork+, CCNA

The lesson for every IT pro is this: the dangerous part of a network is rarely the shiny new gear. It's the seam where new meets old. Engineers who understand both sides of that seam are the ones who keep a country flying.

Quick check: would you have caught it?

A new router at a remote site starts flooding the backbone, and CPUs on older routers across the network spike. Which control is designed to protect a router's own CPU from being overwhelmed?

Spanning Tree PortFast Control Plane Policing (CoPP) HSRP DHCP snooping

Be the engineer who sees it coming.

Our instructor-led CCNA, CCNP and Security courses come with real remote labs, where you can break networks on purpose and learn to fix them.

See upcoming classesTake a free readiness checkPractice free on QuizMe

Sources: FAA, "Nationwide Ground Stop: Questions and Answers"; CBS News; ABC News; ABC7 (L3Harris statement); ABC7 Chicago. Figures differ between sources. Where they do, we've used the FAA's figures or given a range. The investigation is ongoing, and we'll update this post when the FAA publishes its root-cause findings.

More from the blog