Denial-of-service (DOS) attacks increasingly gained reputation over the past few years. As the Internet becomes more ubiquitous, the threat of the denial-of-service attacks becomes more realistic and important for individuals, businesses, governmental organizations, and even countries. There is intensive need to detect an attack in progress as soon as possible. The efficiency of diagnosing the DOS attack using concepts of queuing theory and performance parameter of the system has been investigated in the present work, as the servers definitely have some mechanisms to store and process the requests. Utilizing this concept of queuing theory, the collection of data patterns were generated. With the performance parameter of the system, the analysis of the data pattern had been made to diagnose the network anomaly. Performance analysis and results show the accuracy of the proposed scheme in detecting anomalies.