Bachelor and Master Theses

Background
Smart systems have many connected devices, which makes finding security risks complicated. With the increasing complexity of smart room environments, achieving robust cybersecurity has become a critical challenge. This thesis proposes a novel approach using graph theory to model, identify, and mitigate security threats within smart room systems. By representing smart room components and their interactions as a graph, this thesis will leverage advanced graph-theoretic techniques to detect potential vulnerabilities, strengthening the security framework. While traditional Data Flow Diagrams (DFDs) represent data flows, this proposal adopts a broader graph-theoretic approach to capture more comprehensive aspects of a smart room’s configuration. The graphs will include:

• Physical and logical components (nodes) representing various smart room devices and subsystems.
• Diverse interactions and relationships (edges) detailing data exchanges, permissions, and other connections.
• User interactions and access points that identify points of entry or control within the system.
• Security properties and permissions specific to each component and interaction.

This approach allows for a flexible and detailed representation beyond data flows alone, enabling advanced graph algorithms to be applied for in-depth threat detection and security analysis. While DFDs focus on visualizing data movements, this graph-based model will provide a better depiction of the system’s physical, logical, and security attributes, offering more nuanced insights for safeguarding the smart environment.

Research Questions

1. How can graph theory be applied to model security relationships and interactions in smart room systems?
2. How does a graph-based threat detection approach compare to traditional security methods in terms of accuracy and scalability?

Methodology

1. Graph Representation of Smart Room Systems: This phase will involve mapping the components, data flows, and user interactions within smart rooms as a directed graph. Each device, data source, and user interaction will be represented as nodes, with edges reflecting system interactions such as data exchanges and permissions. The comprehensive representation will incorporate security attributes that detail how components interact within the system.
2. Threat Modeling Using Graph Theory: Graph-based threat vectors will be analyzed through various graph traversal techniques (e.g., breadth-first and depth-first searches) to identify potential pathways for cyber threats. Components of the STRIDE threat modeling framework will be applied to categorize these threats, enabling a structured analysis based on graph data.
3. Evaluation and Comparison: The graph-theoretic approach will be tested in a controlled smart room lab environment to evaluate its performance in threat detection 

Expected Outcomes

• A new way to find security threats in smart room systems using graphs.
• Show that a graph theory-based approach can enhance threat modeling by offering a dynamic, clear view of security structures in smart rooms.
• Ideas for making this method work even better in the future.

• Proficiency in a programming language such as Python, Java, or C++ for implementing graph algorithms.
• Experience with libraries or tools for graph analysis (e.g., NetworkX in Python).
• Basic understanding of IoT or smart room systems, including their components and data flows.
• Basic understanding of cybersecurity principles, including threat modeling and security analysis.

For further information, please contact sara.abbaspour at mdu.se