Traffic flow characterization of cloud applications used in campus network

The paper reports the results from flow-based measurements carried in campus network of AGH-UST. We have conducted detailed analysis of traffic streams in order to find most active online applications. Measurements were carried with the focus to find most active applications, anomalies and also malicious users. It was shown that Facebook, YouTube and Dropbox, they constitute highest and rather stable contributions to whole traffic exchanged with Internet. Presented results show sizes, lengths, inter-arrival times supplemented by average packet sizes of flows belonging to analysed applications.