Lifecycle Cyber-Resilience for Communication and IoT Systems in High-Risk Environments: A Simulation-Validated Architecture for Supply-Chain Assurance, Secure Access Governance, and Explainable Adaptive Operation

In the modern world, the devices that keep our lights on, our factories running, and our hospitals connected are no longer just isolated machines. They are part of vast, interconnected networks where a single weak link can compromise an entire system. This reality has given rise to a field of engineering focused on cyber-resilience: the ability of a system not just to resist an attack, but to anticipate trouble, withstand it, recover quickly, and adapt safely when conditions change. For these systems to be truly secure, protection cannot stop at the moment a device is turned on. Security must follow the device from the factory floor, through its daily use, and into its eventual retirement. This means checking where a device came from, ensuring the person controlling it is who they say they are, and making sure the device can react intelligently when it faces stress or danger.

Cedric Alexander Thornvale, a researcher at DeVry University, has proposed a new way to manage this complex journey. In a recent study, he introduced a framework called the Lifecycle Cyber-Resilience Architecture, or LCRA. This is not a single tool, but a three-part system designed to work together. The first part checks the device's history and health before it even joins the network. The second part acts as a strict gatekeeper, verifying that only the right people can perform sensitive actions. The third part allows the device to change its behavior on the fly to stay safe without shutting down completely. To see if this idea actually works, Thornvale did not just draw diagrams; he built a massive, detailed computer simulation. He created a virtual world containing 6,000 device-assurance records, 8,000 access attempts, and 11,000 end-to-end lifecycle sessions, ranging from normal daily use to severe attacks and environmental stress. The goal was to see if connecting these three layers of protection would stop more threats than using them separately.

The simulation began by testing the first layer: the check on a device's origins. In the real world, a device might be compromised by a dishonest supplier or infected with bad software before it ever leaves the factory. Traditional security often relies on a simple check, like verifying a digital fingerprint of the software. Thornvale's system, however, looks at a much wider picture. It gathers evidence about who made the device, the state of its software, known weaknesses, the reputation of the supplier, and whether the device is still supported. By weighing all these factors together, the system assigns a health score to every device. In the simulation, this multi-evidence approach was remarkably effective. It correctly identified 99.9% of the compromised devices in a standard testing environment, whereas the simple fingerprint check only caught about 62%. The trade-off was that the new system was slightly more cautious, occasionally flagging a safe device as suspicious, but it caught nearly every threat that the old method missed.

Once a device passed the initial check, it faced the second layer: access control. This is where the system decides who is allowed to talk to the device and what they are allowed to do. The simulation tested a scenario where an attacker tried to sneak in by reusing old login credentials or by pretending to have more power than they actually did. The new architecture required that every request be checked against the device's current health score. If a device was flagged as risky, even a legitimate user with a valid password would be blocked from making sensitive changes. The results were decisive. The new system rejected every single modeled attempt by an attacker to gain unauthorized access. In contrast, a system that only checked passwords and ignored the device's health status failed to stop many of these attacks. The study showed that knowing the device is safe is just as important as knowing the user is who they claim to be.

The final layer of the system deals with the device's behavior while it is running. In high-risk environments, devices often face sudden spikes in energy demand, network delays, or strange activity that suggests an attack. Older systems might simply shut down to stay safe, or they might ignore the problem until it is too late. The LCRA system uses a set of clear, understandable rules to decide how to react. It can choose to save energy, monitor the situation more closely, limit its services, or isolate itself completely. Crucially, this decision is not made in a vacuum; it takes into account the device's health score from the first layer and the access rules from the second. When the simulation introduced difficult conditions, this adaptive system kept the service running 94% of the time while successfully containing the threat. A simpler system that used fixed rules, without this flexibility, managed to contain the threat less than 67% of the time and often forced the device to shut down unnecessarily.

The most significant finding of the study came when the researchers looked at the entire system working together. They compared the full three-part architecture against versions where one part was removed or disconnected. When the system could not share information between the layers—when the access gatekeeper didn't know the device was sick, or when the adaptive rules didn't know who was asking for access—the system failed much more often. In the standard testing environment, the full, connected system stopped 98.7% of all malicious incidents from start to finish. When they removed the link between the device's history and the rest of the system, the success rate dropped to 78%. This proved that the value of the architecture lies in the connection between the parts, not just the parts themselves.

However, the study also revealed a difficult reality about security. When the researchers made the simulation more stressful, introducing noisier data and harsher conditions, the system became even better at stopping attacks, catching 99.9% of them. But this increased safety came with a cost. The system began to block legitimate requests more often, mistakenly restricting normal users 18.2% of the time. This highlights a fundamental tension in cyber-resilience: making a system extremely safe often makes it less available. The researchers noted that in the real world, this would require careful tuning to find the right balance, ensuring that the system remains secure without becoming so strict that it stops working for the people who need it.

Throughout the simulation, the researchers also ensured that every decision the system made could be explained. They built a feature that could show exactly why a device was blocked or why a specific action was taken, listing the factors that mattered most. In the tests, this explanation system was highly stable, meaning it gave consistent reasons for its decisions even when the situation changed slightly. This transparency is vital for trust, allowing human operators to understand the machine's logic rather than treating it as a black box.

The study concludes that while the architecture is not a perfect solution for every possible threat, it demonstrates a clear path forward. By linking the history of a device, the identity of its users, and its real-time behavior, organizations can build systems that are far more resilient than those relying on isolated checks. The work was conducted entirely in a simulated environment, meaning the results are a strong proof of concept rather than a final product ready for deployment. The researchers suggest that the next step is to test these ideas on real hardware and with real-world data. Until then, the simulation offers a compelling vision of how security can evolve from a series of static checkpoints into a continuous, adaptive conversation that keeps our connected world safe.