Abstract
Network anomaly detection is essential for rapid fault recovery and increasing network reliability. The paper presents results of the utilization of the digital signatures of network segments (DSNS) for network servers to characterize the traffic that flows through the network servers. An alarm system model is also presented that uses the network servers' traffic characterization generated by DSNS. The obtained results validate the experiment and show, in practice, significant advantages in network management.