Visualization of Network Traffic on Purdue High Performance Computing Resources
MaKayla McCartan, Akash Ravi, Erik Gough
Purdue University is home to several high performance computing (HPC) resources, including campus computing clusters, storage systems and Anvil, a $10M NSF funded supercomputer. These HPC resources are connected to a “Science DMZ” network designed to provide a friction-free path supporting low latency, high-speed data transfer. A Zeek-based intrusion and detection system called PULSAR (Purdue Live Security Analyzer) is used for network monitoring of the Science DMZ. The IDS processes and stores JSON logs at a rate of thousands of events per second. In this work, we use the SIEM to produce visualizations of network traffic on the Science DMZ, showing interesting traffic and attack trends for Purdue's HPC resources.