Key Takeaways

  • Traffic originating at FAA facilities in Arkansas and Texas disrupted communications and radar feeds at 17 of 22 air-route centers.
  • Investigators are examining a Juniper Networks router, its configuration, and its installation during an L3Harris Technologies network upgrade.
  • The incident exposes fault-containment and backup-design risks in interconnected critical infrastructure.

The Federal Aviation Administration has isolated facilities in El Dorado, Arkansas, and Amarillo, Texas, after excessive network traffic destabilized the telecommunications infrastructure supporting US air traffic control.

According to the FAA, the sites generated roughly 1,000 connection attempts every four to five seconds. The resulting instability affected aircraft communications and radar feeds at 17 of the agency’s 22 Air Route Traffic Control Centers. The FAA responded with a nationwide ground stop lasting from approximately 3:44 p.m. to 8:15 p.m. Eastern on November 13.

More than 3,000 aircraft were already airborne when the ground stop began, and all landed safely. Approximately 12,462 flights were canceled that day, although the FAA has cautioned that not every cancellation was directly attributable to the network failure.

The disruption was broader than cancellations alone. NPR reported that FlightAware recorded nearly 25,000 delays and more than 9,000 cancellations on November 13. Cirium counted over 4,000 additional cancellations the following morning. Delta Air Lines was among the carriers heavily affected as aircraft, crews, and passengers ended up out of position.

At the center of the technical investigation is a recently installed Juniper Networks router at the El Dorado radio site. L3Harris Technologies installed the equipment during a transition from aging copper telecommunications lines to fiber. The transportation secretary said the router was sent to an L3Harris laboratory in Florida for further examination.

That does not establish a defective router as the root cause. Investigators are considering the equipment, installation process, and configuration. Hewlett Packard Enterprise, which owns Juniper Networks, told ABC News that the device operated as configured by the installer and did not act improperly based on the information available at the time.

A rate of 1,000 connection attempts every four to five seconds is not especially remarkable when viewed strictly as traffic volume on an enterprise router. The more important question is what each attempt caused elsewhere in the FAA Telecommunications Infrastructure, or FTI.

Repeated attempts could have triggered control-plane processing, session creation, route recalculation, or distribution to multiple devices. A relatively modest stream can become disruptive when network behavior amplifies it. Officials have not publicly confirmed whether BGP, multicast behavior, a Layer 2 loop, Spanning Tree Protocol, or another specific mechanism caused the propagation. Those possibilities remain hypotheses, not findings.

As a precaution, the FAA disabled a broadcasting feature across comparable Juniper Networks routers. Conditions stabilized after the El Dorado radio site and Amarillo radar site were isolated. Authorities had identified no evidence of a cyberattack as of the cited reporting, pointing the investigation toward operational failure, configuration control, and network resilience.

The physical locations may look peripheral. Logically, they were not. FTI is a private network connecting air-traffic centers, towers, radar facilities, and other operational sites. A remote facility can therefore occupy a consequential position in the national architecture, regardless of its size or geography.

There is also a redundancy question. Reported limitations included backup communications channels that used the same underlying telecommunications network. Such an arrangement can provide protection against individual component failures while remaining exposed to a shared network event. Redundancy on paper is less useful when primary and backup paths share a common failure domain.

Even after departures resumed, the transportation system needed time to recover. CNBC reported continuing cancellations and delays as airlines rebuilt schedules. Fixing a router does not instantly return aircraft and crews to their planned airports. That operational tail can outlast the triggering failure by hours or days.

For infrastructure leaders, the episode puts change control under scrutiny. Did testing reproduce production traffic patterns? Were automated rate limits or segmentation controls available? How quickly could operators identify and disconnect the originating sites?

The FAA’s modernization work may reduce risks associated with aging copper links, but replacing old technology introduces transition risk of its own. Staged deployments, independent backup paths, tightly scoped broadcasting functions, rehearsed rollback procedures, and monitoring tied to logical topology can help reduce that exposure. The larger lesson is blunt: in a highly connected operational network, the blast radius of a local change may be national.