Digital signature to help network management using principal component analysis and K-means clustering

The complexity of a network nowadays and its increasingly amount of traffic data has contributed to the occurrence of problems and anomalies. A traffic characterization, called Digital Signature for Network Segment using Flow Analysis (DSNSF) is important to help Network Management in avoiding these problems. We propose two methods to generate a digital signature capable of describing the traffic behavior. For this purpose, we used the statistical method Principal Component Analysis (PCA) and the clustering algorithm K-Means. The resulting DSNSFs are then submitted to testing with real data to evaluate its precision.