CICI: TCR: A Unified Monitoring Approach to Enhancing the Security and Resiliency of Hazard Workflows on Heterogeneous Infrastructures
Funding source: NSF CICI-2319944. Period of performance: 10/01/2023 -- 09/30/2026.Project Overview
Climate change and natural hazards present significant dangers to both humanity and critical cyberinfrastructures (CI). In response, the Earth Sciences community has adopted a cross-boundary approach to monitoring and mitigating these threats by developing CI and workflows that encompass sensors, edge computing nodes, network backbones, and remote cloud data centers. Unfortunately, this cross-boundary approach is vulnerable to both physical (resiliency) and cyber (security) threats. Current state-of-the-art techniques for addressing these threats take a fragmented approach, focusing on specific resources and proving ineffective for the overall hazard workflow. This project proposes the HazardMon framework, which scales horizontally (from sensors to edge cloud to WAN backbones to remote cloud) and vertically (across all layers of the hardware and software stack) to combat these two threats. The project will open-source HazardMon, and results will be disseminated to the community. The broadening participation activities in this project will catalyze community engagement and experiential learning, resulting in a globally competitive workforce. The project will also collaborate with stakeholders, particularly the NSF SAGE project, to facilitate the practical implementation of these ideas.
HazardMon will adopt an interdisciplinary approach that encompasses computer and earth sciences across Rutgers and the University of Oregon, addressing the two threats through three main thrusts. First, HazardMon will utilize the NSF-funded Parasol micro-datacenter to identify threat telemetry for various representative hazard scenarios. Additionally, it will include a monitoring service to observe diverse sensors, WAN backbones, and remote cloud/HPC backends based on the identified telemetry. Second, HazardMon will focus on intelligent techniques to filter and process threat telemetry at the edge, distinguishing between threat telemetry and hazard workflow data. Third, HazardMon will incorporate workflow-specific configuration spaces to develop resilience strategies that mitigate physical threats across CI resources and workflows. Furthermore, it will feature a three-layered mitigation solution to address cyber threats.
People
- Lead PI: Sudarsun Kannan
- Co-PIs: Ram Durairajan (UO PI), Allen Malony (UO), Doug Toomey (UO), Shiqing Ma (UMass), Thu Nguyen (Rutgers), Ulrich Kremer (Rutgers)
- Ph.D. Students: TBD
- B.S. Students: TBD
Publications
- Leveraging Prefix Structure to Detect Volumetric DDoS Attack Signatures with Programmable Switches
Chris Misa, Ramakrishnan Durairajan, Arpit Gupta, Reza Rejaie and Walter Willinger
In IEEE Symposium on Security and Privacy (S&P) (Oakland '24), San Francisco, CA, May 2024.
[PAPER] [CODE]