Security-Informed Safety Case Approach to Analysing MILS Systems

Kateryna Netkachova, Kevin Mueller, Michael Paulitsch, Robin BloomfieldMILS@HiPEAC 2015 · DOI

TL;DR

This paper introduces a "security-informed safety case" approach to analyze and justify the safety of complex systems, particularly those with high security requirements like MILS (Multiple Independent Levels of Security) systems. By integrating security considerations into traditional safety case methodologies, the authors demonstrate how to build more robust and convincing arguments for system dependability, using a MILS-based security gateway as a practical example.

What Problem Does This Paper Address?

Traditionally, safety and security have been treated as separate disciplines in the development of high-assurance systems. However, as systems become more interconnected and open, this separation is no longer tenable. Security breaches can directly impact safety, and vice-versa. This paper addresses the challenge of effectively integrating safety and security analyses into a unified assurance argument. Specifically, it tackles how to create a "security-informed safety case" that not only justifies the safety of a system but also explicitly considers and mitigates security threats and vulnerabilities, especially in the context of complex architectures like MILS, which are increasingly used in critical domains such as avionics.

Methods

The core of the paper is the proposed "Layered Assurance Approach" for constructing security-informed safety cases. This approach builds upon the established Claims, Arguments, and Evidence (CAE) framework, which is a structured way to present engineering justifications. The novelty lies in extending this framework to explicitly incorporate security considerations at multiple levels of abstraction:

  • L0: Policy and Requirements: This highest level defines the system's security and safety policies and their interactions. It involves analyzing how different threat levels and operational modes might necessitate adjustments to these policies.
  • L1: Architectural Layer: This level focuses on analyzing the system's components and architecture, identifying how they contribute to both safety and security. Techniques like guideword-based analysis (similar to HAZOP) and threat modeling (STRIDE) are applied here.
  • L2: Implementation Layer: This is the most detailed level, scrutinizing the implementation of specific components and their integration. It involves examining configuration files, binary objects, and the overall deployment process to ensure that security mechanisms are correctly implemented and enforced.

The authors also leverage the concept of "compositional assurance," where the properties of individual components are used to build arguments for the overall system's dependability, a key principle in MILS architectures.

Datasets

Not discussed in the paper. The paper focuses on a conceptual approach and a case study rather than empirical evaluation on large datasets.

Key Insights and Ideas

  • Interdependence of Safety and Security: The paper strongly emphasizes that safety and security are not independent but deeply intertwined. A system that is not secure is unlikely to be safe, and vice-versa.
  • Layered Assurance is Crucial: A single-level analysis is insufficient. By breaking down the assurance argument into distinct layers (policy, architecture, implementation), it becomes more manageable and allows for a more thorough examination of how safety and security are achieved at each stage.
  • MILS as a Foundation: MILS architectures, with their inherent properties of separation and controlled information flow, provide a strong foundation for building secure and safe systems. However, the paper highlights that MILS alone is not a silver bullet; explicit security-informed safety case analysis is still required.
  • Beyond Traditional Safety: The approach necessitates a shift in thinking for safety engineers, requiring them to consider a broader range of threats, vulnerabilities, and attack vectors that are typically the domain of security professionals.
  • Lifecycle Considerations: The paper acknowledges that the assurance process must extend beyond initial design and implementation to include considerations for adaptation, updates, and maintenance throughout the system's lifecycle, especially as threat landscapes evolve.

Key Results

The paper presents a conceptual framework and demonstrates its application to a MILS-based security gateway. While not providing quantitative results in terms of metrics like error reduction or assurance levels, the key qualitative result is the demonstration that the proposed security-informed safety case approach:

  • Provides a structured and comprehensive way to analyze and justify the combined safety and security properties of complex systems.
  • Facilitates better communication and understanding of interdependencies between safety and security concerns among stakeholders.
  • Highlights specific areas that require more scrutiny when integrating security into safety arguments, such as supply chain integrity and the management of different operational modes.

Limitations and Future Work

  • Formalization: The authors acknowledge that while the CAE structure provides a good framework, further formalization of the reasoning within the security-informed safety cases is needed. This includes linking them to formal models to provide stronger mathematical guarantees.
  • Compositionality and Traceability: More detailed work is required to address issues related to compositionality (how component assurances combine) and traceability between the different assurance layers.
  • Semantics: The paper suggests that a more explicit model-based approach might be necessary to provide a formal semantics for the claim structures, ensuring that lower-level properties truly combine to satisfy top-level claims.
  • Broader Application: While demonstrated on a security gateway, the approach's applicability to a wider range of MILS systems and other critical systems could be further explored.

Summary updated

BibTeX
@inproceedings{netkachova2015security,
  author = {Kateryna Netkachova and Kevin Mueller and Michael Paulitsch and Robin Bloomfield},
  title = {{Security-Informed Safety Case Approach to Analysing MILS Systems}},
  booktitle = {MILS@HiPEAC},
  year = {2015},
  doi = {10.5281/zenodo.47987},
}
Figure