Visualisation of Network Traffic using Dynamic Co-occurrence Matrices

Traffic visualisation is important in several areas e.g. network planning and monitoring, network traffic analysis and intrusion detection. A novelty in the work we present in this paper is the use of texture analysis methods from the domain of digital image processing for network traffic visualisation. We use strategies based on co-occurrence matrices to derive statistical properties for network traffic visualisation and anomalous traffic detection. Based on the fact that some of the statistical properties are related to a certain kind of traffic, which is also reflected in the allocation of the dynamic co-occurrence matrix, we are able to display the global status of our network and show periods, where the traffic behaviour is unusual. Further, we introduce a new parameter, network traffic homogeneity (NTH) as a measure of the local roughness of the network traffic.