Modern digital ecosystems—ranging from the Android software stack to satellite constellations and blockchain networks—are too heterogeneous to be secured as monolithic entities. When a system is composed of millions of lines of code, proprietary hardware, and human operators, a single "security score" becomes meaningless. To address this, researchers have converged on a strategy of hierarchical security decomposition: breaking complex systems into functional, architectural, or administrative layers to systematically map vulnerabilities and prioritize defenses.
The Architecture of a Vulnerability
The core intuition behind this research direction is that a vulnerability's impact and its remediation strategy are dictated by its location in the system's hierarchy. Instead of treating every bug as a generic entry in a database, researchers like Sadegh Farhang and Mehmet Bahadir Kirdan use directory-based mapping. By associating the file paths of patched source code with a predefined system hierarchy, they can visualize which architectural layers (e.g., the Linux kernel vs. the Android application layer) are most prone to "patch gaps"—the delay between an upstream fix and a downstream deployment Hey Google, What Exactly Do Your Security Patches Tell Us?.
This same logic applies to decentralized systems. Daoyuan Wu demonstrates that by mapping diverse blockchain implementations to a standardized four-layer model (application, smart contract, consensus, and network), researchers can identify "hotspots" where specific bug patterns recur across different projects Diving Into Blockchain's Weaknesses.
Functional Segmentation and the Blast Radius
Decomposition is not just about organizing code; it is about limiting the "blast radius" of an attack. In high-stakes environments like Industrial Control Systems (ICS) or satellite networks, researchers use functional segmentation to prevent lateral movement.
- Satellite Systems: Roy Peled and Asaf Shabtai partition the space ecosystem into ground, space, communication, and user layers. This allows them to extend frameworks like MITRE ATT&CK to the space domain, identifying how a breach in a ground station might (or might not) propagate to a satellite's orbital controls Evaluating the Security of Satellite Systems.
- Industrial Networks: Shihao Wu and Quan Pan advocate for dividing industrial environments into distinct zones (Business, Monitoring, Control, Field) connected by controlled conduits. This ensures that a compromise on the plant floor does not grant automatic access to sensitive control zones Network Security in the Industrial Control System.
By treating the system as a set of nested, isolated containers, security professionals can apply the principle of Zero Trust. As explored in Zero Trust for Cyber Resilience, this approach uses micro-segmentation to enforce least-privilege access at every layer, ensuring that trust is never assumed, only verified strategically.
Quantifying the "Invisible" Layers
A recurring theme across this literature is that technical vulnerabilities are only half the story. Jeonghwan Lee and Byungcho Kim apply Socio-Technical System (STS) theory to show that "invisible" administrative security—policies, training, and organizational structure—is often more critical for incident response than "visible" technical tools Incident Response Competence by The Security Types of Firms.
To bridge the gap between these qualitative administrative goals and quantitative technical controls, Jing Yuan utilizes the Analytic Hierarchy Process (AHP). This mathematical framework decomposes high-level protection goals into hundreds of granular technical controls, allowing organizations to calculate a unified numerical score for their security posture Building Security Assessment Index System on Cloud Computing Platform.
Emerging Frontiers: AI and Autonomous Agents
The shift toward autonomous systems—such as LLM-powered agents and self-driving cars—introduces new layers to the hierarchy. Mohamed Amine Ferrag and Leandros Maglaras have proposed threat models that organize vulnerabilities in AI ecosystems from the "prompt" level down to the "protocol" level From Prompt Injections to Protocol Exploits. Similarly, in the automotive sector, Abdulla Obaid Al Zaabi and Chan Yeob Yeun partition the environment into vehicle, communication, and edge cloud layers to account for the risks introduced by distributed third-party components An Enhanced Conceptual Security Model for Autonomous Vehicles.
In all these cases, the goal remains the same: by decomposing a complex whole into its constituent parts, we can move from a state of "security uncertainty" to a structured, measurable, and resilient defense.