SDN-PANDA: Software-Defined Network Platform for ANomaly Detection Applications

The proliferation of cloud-enabled services has caused an exponential growth in the traffic volume of modern data centres (DCs). An important aspect for the optimal operation of DCs related to the real-time detection of anomalies within the measured traffic volume in order to identify possible threats or challenges that are caused by either malicious or legitimate intent. Therefore in this paper we present SDN-PANDA, a ´pluggable´ software platform that aims to provide centralised administration and experimentation for anomaly detection techniques in Software Defined Data Centres (SDDCs). We present the overall design of the proposed scheme, and illustrate some initial results related to the performance of the current prototype with respect to scalability and basic traffic visualisation. We argue that the introduced platform may facilitate the underlying functional basis for a number of real-time anomaly detection applications and provide the necessary foundations for such algorithms to be easily deployed.