NSF Award
2232864
The University of Memphis (UofM) and Florida International University (FIU) are partnering on project LaSIC (Labeled Security Information Capture) to support NSF’s mission to secure scientific assets. Scientific research relies heavily on cyber assets (networked components and services) that must operate securely to support fundamental scientific research and increase confidence in outcomes and conclusions. Developing robust security requires high-quality, trustworthy datasets to tune, test and validate security models and mechanisms. The research value of such datasets increases greatly if they are labeled, which is essential for developing new machine learning (ML) techniques, and establishing ground truth. However, due to their distributed nature, volume and privacy issues, labeled datasets have been extremely hard to produce and distribute to the security community while respecting user privacy.<br/><br/>LaSIC captures, labels and distributes real traffic from AmLight, a production international research and education network serving several scientific communities, to researchers working on securing scientific data and artifacts. AmLight provides connectivity to numerous science disciplines, including Astronomy, Climate, High-Energy Physics, Genomics, and more. AmLight connects Latin America, the Caribbean, and Africa to the U.S. through multiple submarine optical cable systems, as well as other networks such as ESnet and Internet2 in the continental US. LaSIC uses an existing, unique capture infrastructure and capabilities to record, process, and label datasets, including the ability to capture fine-grained network telemetry through Intel Tofino switches, a custom packet capture tool that supports speeds up to 40G, unsampled flow data derived from raw packet capture, network-wide sampled Netflow data, and alerts from a commercial Intrusion Detection System (IDS) deployed in the network. LaSIC provides multimode, structured, and labeled cybersecurity datasets of networking events that support a wide range of security research. Finally, LaSIC distributes these datasets through various mechanisms according to the sensitivity of the data to protect user privacy.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
