|
摘要
|
I propose two frameworks. One framework combines network threats and network faults (CNTNF). This framework incorporates our previous network threat detection and fault localization research. For previous works, I propose three models—the fast filtering and identification system using an ant agent system to effectively defend against denial of service (DoS), quality of service (QoS) attacks, and QoS fault cases, it is called the unified threat identification and fault localization by using ant colony optimization (ACO) (UTFACO), the ant colony system for distributed detection and identification of distributed denial of service (DDoS), namely the distributed detection and identification ant colony system (DDIACS) and the software fault localization (SFL)/network fault localization (NFL) cases are overcome by the spectrum-based SFL (SSFL) system architecture. Additionally, the CNTNF includes the SSFL method to diagnose network faults and multiple QoS fault cases. For this reason, I design a flexible framework, which can be expanded based on the new features when the threats or faults are found and outperformed. The second framework is for the comparison and analysis of the various countermeasures against threats and faults. I develop the attack and defense for forecast and verification modeling framework (ADFVMF). ADFVMF accelerates the development of CNTNF and assesses its contribution value. The experimental results demonstrate that the aggregate total average (ATAVG) of detection rate (DEC-R), ATAVG of accuracy rate (ACC-R), and ATAVG of duration time (DUR-T) are 84.26 %, 88.03 %, and 11.38 s, respectively. Consequently, CNTNF is a stability framework based on the boundary limitations and the optimization of parameters in terms of efficiency and effectiveness. |