Design and Practice Exploration of Software Security Experimental Teaching

F Xu, Zhenyu Zhu, Feng Gao, Xie Xiao, Yaojie ChenAtlantis 2024 · DOI · OpenAlex

Abstract

In order to cultivate students' comprehensive ability and innovative thinking, meet the requirements of engineering practice ability under the new engineering background, and solve the difficulties faced by software experiment teaching at present, combining the advantages of traditional teaching, this paper puts forward new ideas and new ideas for software safety experiment teaching design and practice. The combination of independent experiment and comprehensive experiment, classification teaching, the introduction of competition mechanism heuristic teaching, using online experimental teaching platform, focusing on process assessment and other aspects are implemented into the experimental teaching process. Finally, the subsequent teaching improvement suggestions are put forward.

Introduction

Software security is a high-stakes field, yet traditional education often fails to prepare students for its complexity. Common pitfalls include passive learning, where students treat labs as checklists, and the "tool usage trap," where students rely on automated tools without understanding underlying mechanisms like memory management. This paper, originally presented at ITEI 2024, outlines a modernized framework implemented over five years to transform students from tool users into security analysts.

Methodology

The authors implemented a four-pronged strategy to bridge the gap between theoretical foundations and practical exploitation/defense.

Categorized Teaching

Students are divided into groups (Basic, Intermediate, Advanced) and tracks (Instructional vs. Competition). This allows for personalized guidance; for instance, advanced groups tackle "very challenging" tasks like malware family classification while basic groups focus on foundational secure coding.

Categorized teaching model showing the relationship between student groups, guidance levels, and experiment difficulty.

Hybrid Experiment Design

The curriculum blends "Independent Experiments" focused on specific concepts like SQL injection with "Comprehensive Experiments" that integrate multiple skills. A key principle is "loose coupling," ensuring that failure in a preliminary step doesn't stall progress in complex subsequent tasks.

Classification of experiment types including innovative, competition, and practical application experiments.

Competition-Heuristic Model

The course introduces competitive mechanisms where students research hot topics and present findings. Students are encouraged to analyze vulnerabilities in real-world IoT or mobile devices for accumulated points, fostering initiative through a public leaderboard.

Process-Oriented Assessment

The grading rubric prioritizes active participation over final reports. The final grade is split: 10% attendance, 70% performance during experiments (analytical ability and defense scores), and 30% for the written report.

Implementation and Key Insights

The curriculum utilizes the Istudy online platform for sandboxed virtual laboratory experiments. Key experimental scenarios include:

  • Vulnerability Analysis: Buffer overflows, SQL injection, XSS, and "Double Free" experiments.
  • Malware Analysis: Behavioral analysis using samples like "Panda Burning Incense."

A major focus is placed on reverse-engineering the "why." For example, in buffer overflow labs, students must explain the memory management and compiler implementation details that make the attack possible, rather than just triggering an exploit.

Results and Future Work

After five years of implementation, the authors report marked improvements in autonomous learning and a reduction in student frustration due to the categorized model. The 70% performance-based grading provided a more accurate reflection of practical capability than traditional reports.

Future work will focus on deeper industry integration, inviting professionals from companies like Qihoo 360 and Tencent as mentors, and incorporating cutting-edge content such as blockchain security and machine learning applications in threat detection.

Summary updated

Abstract

In order to cultivate students' comprehensive ability and innovative thinking, meet the requirements of engineering practice ability under the new engineering background, and solve the difficulties faced by software experiment teaching at present, combining the advantages of traditional teaching, this paper puts forward new ideas and new ideas for software safety experiment teaching design and practice.The combination of independent experiment and comprehensive experiment, classification teaching, the introduction of competition mechanism heuristic teaching, using online experimental teaching platform, focusing on process assessment and other aspects are implemented into the experimental teaching process.Finally, the subsequent teaching improvement suggestions are put forward.
BibTeX
@inproceedings{xu2024design,
  author = {F Xu and Zhenyu Zhu and Feng Gao and Xie Xiao and Yaojie Chen},
  title = {{Design and Practice Exploration of Software Security Experimental Teaching}},
  booktitle = {Atlantis},
  year = {2024},
  doi = {10.2991/978-94-6463-560-7\_31},
}
Figure