一句话：提出五种变异调度策略和三种种子语料选择策略，提升 SAP HANA 单元级持续模糊测试的可配置性与效率。
本文报告了在 SAP HANA 组件上进行有效持续模糊测试的工业经验，核心问题是现有单元级模糊测试框架缺乏可配置性，难以针对不同组件有效使用语法感知变异器并优先检查变更代码。为此，作者提出五种新的变异调度策略，以更有效地利用语法感知变异器，并设计三种新的种子语料选择策略，用于配置模糊测试活动以优先检查变更代码。这些扩展被集成到单元级模糊测试框架中。实证结果表明，该扩展使用户能够针对每个目标组件专门配置框架，从而有机会提升模糊测试的有效性和效率。
This paper presents industrial experiences on enhancing the configurability of a fuzzing framework for effective continuous fuzzing of the SAP HANA components. We propose five new mutation scheduling strategies for effective uses of grammar-aware mutators in the unit-level fuzzing framework, and three new seed corpus selection strategies to configure a fuzzing campaign to check on changed code in priority. The empirical results show that the proposed extension gives users chances to improve fuzzing effectiveness and efficiency by configuring the framework specifically for each target component.
DOI 原文 ·
@inproceedings{paperbot6499,
title = {Improving Configurability of Unit-level Continuous Fuzzing: An Industrial Case Study with SAP HANA},
author = {Hanyoung Yoo and Jingun Hong and Lucas Bader and Dong Won Hwang and Shin Hong},
booktitle = {ASE},
year = {2021},
doi = {10.1109/ase51524.2021.9678685}
}