Abnormal traffic detection method for customer service cloud service data center based on improved Transformer
Lihua Gong, Dongze Wang, Xia Zhang, Haiyang Ren, Ruidong Fu, Yang Liu · 2025
In the cloud computing environment, customer service cloud service data centers face a large amount of data traffic, which may hide abnormal behaviors such as network attacks and data leaks. At present, abnormal traffic detection methods may be difficult to adapt to this big data environment due to issues such as low detection accuracy and efficiency. Therefore, this paper proposes a method for detecting abnormal traffic in customer service cloud service data centers based on an improved Transformer. Extracting traffic characteristics from customer service cloud service data centers in cloud computing, constructing an improved Transformer based abnormal traffic detection model for customer service cloud service data centers, and solving the problem of sequence to sequence detection. On this basis, differential evolution algorithm is introduced to optimize the abnormal traffic detection model of customer service cloud service data center, achieving abnormal traffic detection of customer service cloud service data center. The experimental results show that the average MAE value of the proposed method is below 4.0, and the abnormal traffic detection time is only 66ms, which can effectively improve the efficiency and accuracy of data traffic anomaly detection in customer service center.