On Friday, November 13, 2026, the FAA ordered a nationwide ground stop that lasted more than four hours. The FAA has found no sign of a cyberattack. It began with a routine network change: a new router at one remote radio site.
The root cause is still under investigation. What is already on the record is a lesson in change management, and it applies to any network.
What the change did
The FAA’s network service provider, L3Harris, was replacing a router at a radio site in El Dorado, Arkansas. The FAA calls the work routine. Work at major facilities is typically done at night. This site is remote, so the replacement was done in the middle of the day.
Shortly after the new router went in, the network was flooded: about 1,000 connection attempts every four to five seconds, or 200 to 250 a second. The FAA’s initial investigation points to the El Dorado site and to a radar site in Amarillo, Texas. The FAA says its legacy TDM equipment became unstable under that load.
Communications and radar feeds were disrupted at 17 of the FAA’s 22 en route centers. Controllers in Indianapolis noticed first, as static on their radio links to pilots.
The fix was isolation. The provider blocked the two sites, and the network stabilized. As a precaution, the FAA then disabled a broadcasting function on the new routers and on about 60 similar ones. Air traffic resumed at 8:15 p.m. Eastern, nearly five hours after the problem began.
What caused the flood is not settled: the installation, the configuration, or the router itself. HPE, whose Juniper unit made the router, says it worked as intended.
Why it looked safe
The risk was judged by the site, not by its reach. Daytime work at a remote site was assumed to affect few flights. But the router joined a network that en route centers across the country share. The FAA now says it will reevaluate when it does replacement and maintenance work.
Nothing had to fail. If its maker is right, the router did what it was set up to do. The trouble was what it sent into the network around it, and what the oldest equipment on that network could absorb. No single device’s log shows that combination.
The symptom appeared somewhere else. Indianapolis heard static. The source was hundreds of miles away, in southern Arkansas. The fastest way to connect the two is to ask what changed, and to have the answer on record.
What checks before and after a change are for
We do not know which checks were run in El Dorado. The FAA has not said. But this is the kind of failure a change procedure exists to catch.
Before the change. State the scope by reach: what the new device connects to, and what it will send that the old one did not. Write down what must not change on the rest of the network, and the trigger for rolling back.
After the change. Check the network around the device, not only the device: its neighbors, the traffic on shared circuits, and the services that ride on them. Compare each with the state captured before the change. Then check again a few minutes later.
When something looks wrong. Ask what changed first. Here the answer was a router replacement, and isolating the sites involved was the fix. With a rollback trigger agreed in advance, that is the first move.
Our free change verification checklist lists the before-and-after checks by change type.
Where REAP fits
This is a change-management problem. A proper change procedure, with checks before and after the work, would have helped isolate this problem sooner, or roll the change back. Raja Tadimeti, REAP’s CEO, made that case at ONUG Fall 2026 in his talk The Change Succeeded. Did the Network? The slides, the talk in writing, and its fifteen-item checklist are on that page.
REAP does two things for a change. Before the window, it assesses the proposed change against topology, reachability, security policy, and known CVEs. That is available today. After the window, it compares configuration and state with what was captured before, and puts the evidence in the change ticket. We deliver that in pilots today, on one real change window with your team. See Change Assurance.
What a network team can take from it
A change is not finished when the device comes up. It is finished when the network around it has been checked. Restoring service and establishing the cause are two different jobs, and the second needs evidence collected at the time: what changed, what the network looked like before and after, and which paths were affected.
A note on the facts. The account of the ground stop comes from the FAA’s questions and answers on the ground stop, published November 15, 2026, and from reporting through that date by the Wall Street Journal, as republished by To Vima, and KATV in Little Rock. The figure of 200 to 250 connection attempts a second is our arithmetic on the FAA’s number. The FAA’s investigation is open, and it has not published findings. REAP was not on the FAA’s network and has no part in the investigation. We will update this post when the findings are public.