CAJ | 학술논문

在国际前沿技术 EFS(evolutionary fuzzing system)的研究基础上,提出基于控制流序位比对算法的智能Fuzzing 测试方法.根据遗传算法的内在属性演算得到基于序列比对的适应度函数,并有效地计算出需要搜索的程序逻辑空间.最后给出了该方法与2种传统 Fuzzing 方法的测试性能的实验结果比对,证明了该方法能够充分利用遗传算法属性中并行性进行智能地程序逻辑学习,具有逻辑覆盖面广、搜索导向性强的优点,能够提高漏洞挖掘能力.
재국제전연기술 EFS(evolutionary fuzzing system)적연구기출상,제출기우공제류서위비대산법적지능Fuzzing 측시방법.근거유전산법적내재속성연산득도기우서렬비대적괄응도함수,병유효지계산출수요수색적정서라집공간.최후급출료해방법여2충전통 Fuzzing 방법적측시성능적실험결과비대,증명료해방법능구충분이용유전산법속성중병행성진행지능지정서라집학습,구유라집복개면엄、수색도향성강적우점,능구제고루동알굴능력.
Flowing the way introduced in the research of evolutionary fuzzing system (EFS), a smart fuzzing method was proposed based on the node comparison algorithm among the control flow sequences. Through mapping program execu-tion flow sequences onto the control flow sequences, the isomorphism relationship between dada search space and pro-gram logic space was established. The analyzed results prove that the method is capable of mining a mass of information from group data effectively, and is able to fully utilize the parallelism of genetic algorithm to guide the fuzzing test.