Understanding and Managing Risks on a Network using a Graphical Model of Connection Data 

Cloud based computing networks and storage systems offer tremendous advantages to their clients. These include huge cost reduction and incredible scalability. However, these benefits come with their own risk, that is security.  Securing the cloud against attacks has been an open research problem ever since these systems have been in use.  In this project, we propose a methodology in modeling the network vulnerabilities as possible precursors to threats based on graphical formulations of the network which depicts entities as nodes with calculated features and has edges representing their connection. The objective of an intrusion detection system (IDS) is to detect these attacks and identify their sources in real time. These will be done based on a graphical signal processing formulation of multiple sparse measurements of the edge of the network.

Collaborators: David Maluf and Dan Tan

Current Student Researcher: Arda Bayer

Funded By: