Machine Learning-Based Detection of Synonymous IP Flood Attacks on Server Infrastructure

In this research work, five different machine learning techniques, namely Random Forest, SVM, GBM, CNN, and IsolationForests are compared in identifying synonymous IP flood attacks against the server architecture.Through the analysis of these algorithms with a large number of normal and attack traffic samples, thus improving server environment security and protection capabilities against more advanced threats.The performance measures used were F1 Score, Precision, Recall, Accuracy, and AUC-ROC (Area Under the Receiver Operating Characteristic Curve).The outcome was that Random Forest and GBM models were highly accurate, recording F1 Scores of 0. 92 and 0. 50 attacking accuracy] and F1 Score of 0. 93 respectively, while CNN Floyd was also proven to have satisfied exceptionalism as depicted by the F1 Score of 0. 94.SVM and Isolation Forests also were at the same level revealing F1 Scores equaled to 0. 88 and 0. 90 respectively.Accordingly, these findings support the use of machine learning methods for enhancing the real-time identification and counteraction to IP flood attacks.In the end, the sector's strengths and weaknesses of each algorithm are identified and replicated, which was followed by the recommendation of expanding the future research in developing the methods of the hybrid model or ensemble method that would improve the detection accuracy and adapt to the dynamic cyber threat environment.