IGEL Blog

The FAA Disruption Is a Warning: America Must Engineer for Cyber Survivability
Co-authored by John Walsh and Randy Resnick
The disruption of U.S. air traffic operations is a warning for every leader responsible for critical infrastructure. Whether the immediate cause is a cyberattack, software fault, or communications failure, the strategic question is the same: How can a localized failure threaten an interconnected national service, and what happens when the next disruption is deliberate?
Grounding aircraft may be the appropriate safety decision when essential systems are unavailable. But a safe shutdown is not evidence of a resilient architecture. The more demanding standard is whether critical services can continue safely in a degraded state, with failures contained and recovery assured. The incident alone does not establish how U.S. infrastructure would fare against a coordinated attack. It does expose the stakes of systemic dependence.
The threat has changed faster than the infrastructure
Traditional cybersecurity practices, including identifying vulnerabilities, patching, monitoring, detection, and response, remain essential. But they are no longer sufficient on their own. AI can accelerate reconnaissance, vulnerability discovery, and attack preparation. Agentic systems also introduce new exposure when autonomous workflows are compromised, misconfigured, or given excessive privileges. These capabilities expand the range of actors able to mount consequential attacks, even if sophistication still varies.
Meanwhile, operational technology (OT) in power, water, aviation, healthcare, and manufacturing was often engineered for long service life and deterministic operation, not rapid software change. Patches may require outages, vendor validation, safety assessments, or recertification. An attacker can move faster than an operator can safely update a production system.
We cannot patch our way to resilience, nor can we make detection our only line of defense. Security architecture must reduce exposure before compromise and limit consequences afterward.
Engineer for prevention, containment, and continuity
Critical infrastructure needs three mutually reinforcing capabilities. Prevention means hardened endpoints and workloads, verified device integrity, least-privilege access, and Zero Trust decisions based on continuously evaluated context, not credentials alone. Containment means segmentation, independent security zones, and controls that prevent one compromised endpoint or application from taking down an entire mission. Continuity means essential services remain safe and available through alternate communications, isolated recovery mechanisms, and rehearsed degraded-mode operations.
The priority is to find and eliminate “break one, break all” dependencies. Two backup networks are not truly redundant if they share a control plane, identity provider, power source, or recovery dependency. Independence must be demonstrated, not assumed. Recovery time objectives (RTOs) and recovery point objectives (RPOs) should be defined against mission needs and validated in realistic conditions. Civilian systems supporting national essential functions should adopt the mission-assurance mindset used in other high-consequence environments: assume partial compromise, plan for simultaneous failures, and preserve the mission within explicitly defined safety boundaries.
These principles must also shape procurement and modernization. Architecture reviews should ask whether a proposed platform introduces a shared point of failure, whether emergency operations require cloud connectivity, and whether suppliers can support restoration when enterprise management tools are unavailable. The cheapest design at steady state may be the most expensive during a national emergency.
Operational leaders should define the minimum service that must survive: which functions are indispensable, which can be suspended, and which manual or local procedures are safe. That prioritization determines where to spend on independent power, communications, skilled personnel, spare capacity, and recovery equipment. Without it, organizations may build impressive redundancy around the wrong assets.
Make Zero Trust an operational architecture
Zero Trust is more than authentication and network admission. It must extend to endpoint integrity, workload execution, identity, device posture, segmentation, and access to OT assets. A legitimate credential does not make an untrusted device safe.
NIST SP 800-207, the NIST Cybersecurity Framework, and IEC 62443 offer useful architectural and governance foundations, alongside relevant U.S. defense guidance for IT and OT. Yet compliance is a starting point, not proof that operations will survive an attack. Organizations can protect difficult-to-patch legacy environments through isolation, tightly governed access, and hardened intermediary endpoints rather than assuming immediate replacement is feasible. Immutable endpoint approaches, including those used by IGEL, illustrate one possible layer of that strategy, but cannot substitute for system-wide resilience engineering.
Stress-test the entire mission, not just its components
This is the missing discipline: end-to-end stress testing for resilience and survivability. Component-level penetration tests, backup checks, and tabletop exercises cannot establish whether a service will survive cascading failure. Operators should conduct controlled, safety-approved exercises across the full chain: endpoints, identity, applications, OT control systems, networks, third parties, communications, human procedures, and recovery infrastructure.
Scenarios should combine cyber compromise with loss of primary connectivity, corrupted management systems, unavailable identity services, failed failover, and constrained staffing. Test whether critical functions can operate in degraded mode, whether segmentation contains the blast radius, and whether teams can recover using genuinely independent paths. In safety-critical environments, use digital twins, simulation, and staged exercises where live disruption would create unacceptable risk.
Measure mission outcomes: time to detect and isolate, time to restore minimum viable service, duration of safe degraded operation, and the percentage of critical functions sustained. Record hidden dependencies and repeat the exercises after architectural changes. Survivability is not a claim on a slide; it is a demonstrated capability.
A credible exercise program must include executives, operators, safety engineers, incident responders, suppliers, and public-sector partners. It should establish explicit stop conditions and avoid introducing hazards into live operations. Independent observers can distinguish a scripted demonstration from a genuine test, while after-action reviews should produce funded remediation plans with accountable owners and deadlines.
Stress testing also exposes governance failures: ambiguous authority to disconnect a system, missing contracts for emergency vendor support, or a recovery plan that depends on personnel who cannot reach the facility. Technical resilience and organizational resilience are inseparable.
Treat cyber resilience as a national security imperative
The United States needs coordinated action among federal agencies, standards bodies, infrastructure operators, technology providers, and the defense and intelligence communities. A national-scale initiative should map cross-sector dependencies, accelerate practical Zero Trust adoption across IT and OT, and prioritize independently recoverable communications and essential services.
Public policy should also encourage investment, especially among privately owned infrastructure operators. Options worth evaluating include accelerated first-year expensing for qualified cyber-resilience investments, coupled with measurable outcomes and appropriate safeguards. The ambition should match the interconnected nature of the problem.
Boards and policymakers must stop equating activity with readiness. Patch counts, alerts, and audit completion say little about whether a hospital can treat patients, a water utility can maintain safe service, or an airport can sustain essential operations during a major disruption. The decisive measures are prevention, containment, continuity, and verified recovery.
Our critical infrastructure must be engineered for its worst day, not merely optimized for its best. Cyber resilience is now a national security imperative, and that includes the ability to prove survivability under stress.