Crowdsourcing Approach for Developing Hands-On Experiments in Cybersecurity Education

Le Wang, Zhihong Tian, Zhaoquan Gu, Hui Lu

Abstract

Cybersecurity education requires hands-on experiments that keep pace with rapidly evolving threats. However, a gap exists between the technical skills of engineers and the pedagogical needs of faculty. This paper introduces the Resource Crowdsourcing for Cybersecurity Framework ($RC^2F$), which utilizes incentive models to create a sustainable marketplace for lab development. Validated through a university course, the framework demonstrated high productivity and effective quality control, producing over 80 hours of content in a single semester.

Introduction

Cybersecurity is uniquely characterized as a "post-companion" discipline, where new vulnerabilities and attack vectors emerge almost immediately following software releases. This rapid evolution creates a significant challenge for educators: maintaining hands-on laboratory experiments that reflect current real-world threats. Traditional cyber ranges often suffer from stagnation due to a disconnect between technical experts (engineers) who can build realistic labs and faculty who possess the pedagogical expertise but lack the time or specialized experience for high-volume development.

The Resource Crowdsourcing for Cybersecurity Framework ($RC^2F$)

To address this gap, the authors propose the Resource Crowdsourcing for Cybersecurity Framework ($RC^2F$). This system conceptualizes lab development as a marketplace involving three key stakeholders:

  1. Constructors: Skilled engineers or senior students who develop the experiments.
  2. Demanders: Faculty members who define the educational requirements and curriculum needs.
  3. Users: Students who consume and learn from the developed labs.
The basic crowdsourcing model for cybersecurity experiments, illustrating the relationships between Demanders, Constructors, and Users.

Incentive Mechanisms: FIM and BIM

The sustainability of $RC^2F$ relies on two mathematical incentive models designed to maintain equilibrium within the marketplace:

  • Forward Incentive Model (FIM): This model motivates the Constructors. It calculates rewards based on the quantity of work ($Count_U$), quality scores from users ($Mark$), and financial compensation ($Money$).
  • Backward Incentive Model (BIM): This ensures that the requirements set by Demanders are met, preventing a mismatch where Constructors build technically interesting but pedagogically irrelevant content.
The framework incorporating the Forward Incentive Model (FIM) to drive quality and participation among Constructors.

Experimental Validation

The framework was tested at the Cyberspace Institute of Advanced Technology (CIAT), Guangzhou University, specifically within the "Software Security and Analysis" (SSA) course. A group of 21 senior students acted as constructors, bidding on 12 specific lab requirements, such as Shellcode Programming and Heap Overflows.

Key Results

The implementation yielded several significant insights:

  • High Productivity: The course generated over 80 hours of hands-on content, a volume unattainable by a single faculty member in one semester.
  • The "Saddle" Effect: Participation followed a U-shaped curve, with high initial interest, a dip during the difficult middle phases, and a final surge driven by grade and financial incentives.
  • Quality Correlation: Data confirmed a strong correlation between high user satisfaction scores and financial payouts, validating the FIM's ability to reward quality over mere quantity.

Conclusion and Future Work

The $RC^2F$ framework successfully bridges the gap between technical skill and educational need. By incentivizing senior students to build labs for their juniors, it creates a self-sustaining learning ecosystem. Future work will focus on refining collision rules for duplicate submissions and scaling the model to involve professional security firms as constructors.

Reference: Wang, L., Tian, Z., Gu, Z., & Lu, H. (2019). Crowdsourcing Approach for Developing Hands-On Experiments in Cybersecurity Education. IEEE Access, 7, 169066-169072.

Summary updated

Abstract

The operational skills are essential for the cybersecurity practitioners and the hands-on experiments can be utilized to train their practical skills. As malicious attacks are continually changing and new vulnerability appears quickly with the increasing number of new software, it has been a challenging problem to develop hands-on experiments rapidly and continuously which are consistent with realistic threat scenarios. Normally, skilled engineers are familiar with threat scenarios and the attack/defense skills, but they have no motivation to develop hands-on experiments for education. In contrast, the faculties are good at teaching and training but they have no resource or experience to deploy lots of hands-on experiments by themselves. In this paper, we propose a framework called RC <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> F by adopting the crowdsourcing approach, which acts as a platform for the engineers and the faculties as well as channels between them for resource exchanging. In the framework, we also introduce two incentive models to motivate the participants, where the constructor-oriented Forward Incentive Model (FIM for short) encourages the engineers by the incentive mechanism and the demand-oriented Backward Incentive Model (BIM for short) balances the educational resources for the faculties. We also provide a comprehensive case based on the practical data to analyze the proposed framework. And the results validate that the proposed framework as well as the FIM and BIM models can address the challenge in developing hands-on experiments.
BibTeX
@article{wang2019crowdsourcing,
  author = {Le Wang and Zhihong Tian and Zhaoquan Gu and Hui Lu},
  title = {{Crowdsourcing Approach for Developing Hands-On Experiments in Cybersecurity Education}},
  journal = {IEEE Access},
  year = {2019},
  doi = {10.1109/access.2019.2952585},
}
Figure