TL;DR
Modern ICT systems are expected to last for decades, but current security and safety paradigms are designed for much shorter lifespans and foreseeable threats. This paper introduces "sustainable security & safety" (S3) as a new paradigm to ensure systems remain secure and safe over their entire, often very long, design lifetimes, even against unknown future threats. It outlines the significant challenges and proposes design principles to achieve this crucial goal for long-lived systems.
What Problem Does This Paper Address?
The core problem this paper tackles is the growing mismatch between the expected operational lifetime of many modern ICT systems and the ability of current security and safety engineering practices to maintain their integrity over such extended periods. We're increasingly integrating complex ICT into long-lived assets like cars, critical infrastructure, and smart buildings, which are designed to operate for 25+ years. However, traditional security approaches focus on known threats and adversaries, while safety approaches deal with accidental faults. Neither adequately prepares systems for the unforeseeable threats, evolving attack vectors, and environmental changes that are virtually guaranteed to occur over decades. This leaves these critical systems vulnerable to becoming unsafe or insecure long before their intended end-of-life.
Methods
The paper proposes a new paradigm called "sustainable security & safety" (S3). This paradigm is built on two key characteristics:
- Bridging the Safety-Security Gap: It aims to unify the principles of safety (dealing with accidental faults) and security (dealing with malicious adversaries) into a common framework. This involves addressing both accidental faults and vulnerabilities, as well as malicious attacks and intrusions, under a single body of knowledge.
- Anticipating the Unforeseeable: S3 explicitly acknowledges that long-lived systems will face threats and changes that cannot be predicted at design time. Therefore, it shifts the focus from defending against known threats to building systems that can adapt and remain secure and safe despite encountering unknown future challenges.
The authors then identify and categorize the main challenges to achieving S3, which include subsystem failures (hardware, software, specification), requirement changes (regulatory, user expectations, design lifetime), environmental changes, new threat vectors, unilateral system/service changes, third-party system failures, maintenance resource unavailability, maintainer changes, and ownership changes. Finally, they propose several design principles to address these challenges.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- The "Unforeseeable Horizon": The most critical insight is that for long-lived systems, the "foreseeable horizon" of threats and changes is insufficient. We must design for the "unforeseeable horizon," meaning systems need resilience against threats and conditions that are currently unknown and perhaps even unimaginable.
- Security-Dependability Gap: The paper highlights the existing gap between security and dependability (safety) approaches. Security often assumes an intelligent adversary, while dependability assumes random faults. This dichotomy needs to be reconciled for long-term system integrity.
- Systemic Nature of Failures: Subsystem failures, whether accidental or malicious, must be treated as system faults. This requires mechanisms that can handle failures at the system level, not just at the subsystem level.
- Long-Term Secrets are Hard: Maintaining the confidentiality of secrets over decades is extremely challenging, as leaked data cannot be re-secretized. This necessitates minimizing secret usage and employing advanced techniques like secret sharing.
- Lifecycle Management is Crucial: S3 isn't just about initial design; it's about managing the system throughout its entire lifecycle, including production, maintenance, and even end-of-life, considering changes in maintainers and ownership.
Key Results
This paper is primarily a conceptual and visionary work, outlining a new paradigm and its associated challenges and design principles. It does not present quantitative results from experiments or simulations. Instead, its key "results" are:
- A clear articulation of the Sustainable Security & Safety (S3) paradigm.
- A comprehensive taxonomy of challenges that long-lived ICT systems face in maintaining security and safety over time.
- A set of design principles intended to guide the development of S3-aware systems, including:
- Well-defined Components and Isolation
- Avoiding Single Points of Failure
- Multiple Lines of Defense
- Long-term Secrets and Confidentiality
- Robust Input Handling
- Containing Subsystem Failure
- Replicating to Tolerate Subsystem Failure
- Diversifying Nodes and Components
- Minimizing Assumptions
- Simplicity and Verifiability
Limitations and Future Work
The authors acknowledge that achieving the full vision of sustainable security & safety is a non-trivial challenge and that advances in this direction are needed. They state that "even if the full vision of sustainable security & safety is not fully achieved, any advances in this direction could have a significant impact on the design of future systems."
Future work would involve developing concrete technologies and approaches that implement the proposed design principles. This includes further research into:
- Techniques for anticipating and mitigating unforeseeable threats.
- Robust mechanisms for handling evolving requirements and environmental changes.
- Advanced methods for managing secrets and confidentiality over long periods.
- Architectures that inherently support fault tolerance, adaptation, and secure evolution throughout a system's lifecycle.
- Methods for securely managing changes in maintainers and ownership.
Summary updated