An ontology-based approach to security pattern selection

Hui Guan, Hongji Yang, Jun WangInternational Journal of Automation and Computing 2016 · DOI

Abstract

Software developers often struggle to select appropriate security patterns due to the sheer volume of available options and a lack of standardized classification. This paper presents a semantic framework that utilizes a formal ontology in OWL (Web Ontology Language) to bridge the gap between high-level security requirements and technical implementation patterns. By integrating risk analysis with semantic reasoning, the proposed system enables an automated search engine to navigate complex pattern repositories, providing non-expert developers with precise, context-aware security solutions.

Introduction

Security patterns represent reusable best practices for solving recurring security problems. However, with over 400 documented patterns, developers—particularly those without specialized security expertise—face a "needle in a haystack" problem. Existing classification schemes are often too abstract or inconsistent, leading to error-prone manual selection. This research addresses the challenge by developing a structured Security Knowledge Base that automates the mapping of security requirements to specific patterns.

Methodology

The proposed approach follows a three-stage pipeline designed to transform abstract security needs into concrete design patterns.

Figure 1: High-level process flow for security pattern selection and integration.
  1. Requirement Elicitation: Security requirements are formalized as a 4-tuple: <Asset, Threat, Security Attribute, Priority>. This structure ensures that requirements are grounded in specific risks.
  2. Ontology Development: Using the METHONTOLOGY framework, the authors developed a core security ontology. This is divided into a Security Requirement Sub-ontology (sr), covering assets and vulnerabilities, and a Security Pattern Sub-ontology (sp), which classifies patterns by lifecycle stage, architectural layer, and threat type.
  3. Pattern Search Engine: An automated algorithm (Algorithm 1) utilizes the OWL API to perform semantic reasoning. Unlike keyword-based searches, this engine infers suitability by matching requirement properties to the semantic descriptions of patterns.

The Security Knowledge Base

The core of the framework is the semantic mapping between security threats and pattern capabilities. By utilizing the STRIDE model (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, and Elevation of Privilege), the ontology provides a standardized language for describing security problems.

Figure 4: Conceptual model illustrating the semantic relationships between assets, threats, and security patterns.

As shown in the conceptual model, threats are linked to assets through vulnerabilities, while security patterns are mapped to the specific security problems they solve. This multi-dimensional classification allows the system to consider the architectural layer (e.g., Network vs. Application) and the application domain (e.g., Web and J2EE) during the retrieval process.

Evaluation and Results

The framework was evaluated using a curated repository of 32 seminal security patterns. The prototype system was tested against "competency questions"—real-world queries a developer might pose.

The results demonstrated high precision in context-aware retrieval. For instance, a query regarding protecting sensitive data from network eavesdropping successfully returned specific patterns like "Secure Pipe" and "Secure Communication." By filtering results based on architectural layers and lifecycle stages, the system significantly reduced the "noise" typically encountered in manual searches.

Conclusion and Future Work

This ontology-based approach provides a robust foundation for automating security pattern selection. While the current repository is limited to 32 patterns, the framework is designed for scalability. Future research will focus on aligning the ontology with international standards such as ISO/IEC 27001 and integrating the search engine into broader expert systems to assist developers in the instantiation of patterns within their software models.

Summary updated

BibTeX
@article{guan2016ontology,
  author = {Hui Guan and Hongji Yang and Jun Wang},
  title = {{An ontology-based approach to security pattern selection}},
  journal = {International Journal of Automation and Computing},
  year = {2016},
  doi = {10.1007/s11633-016-0950-1},
}
Figure