Uploaded October 2026 | Updated November 2026, 3 days ago
This project seeks to coordinate network traffic among agentic and scientific workflows across institutional boundaries. As DOE and other research organizations move toward AI-driven autonomous laboratories where robotics, edge AI, real-time analysis and intelligent feedback must operate seamlessly across distant locations, existing passive scientific networks have become bottlenecks that interfere with reliability and resilience. Distributed AI and agentic workflows need adaptive routing to prevent congestion, enable GPU-to-GPU transfers and take full advantage of 6G and accelerated network technology. SciNet, a science-aware platform for research network operations, will curate multimodal network operations data and tag packets with science project information to allow correlation of data from heterogeneous sources, handle sensitive data securely and anticipate network requirements to deliver shorter experimental latency, fewer workflow failures and faster discovery cycles across science domains to satisfy workflow needs.
To learn more about this project, visit ornl.gov/projects/ai-powered-operations-self-driving-science-research
This project seeks to coordinate network traffic among agentic and scientific workflows across institutional boundaries. As DOE and other research organizations move toward AI-driven autonomous laboratories where robotics, edge AI, real-time analysis and intelligent feedback must operate seamlessly across distant locations, existing passive scientific networks have become bottlenecks that interfere with reliability and resilience. Distributed AI and agentic workflows need adaptive routing to prevent congestion, enable GPU-to-GPU transfers and take full advantage of 6G and accelerated network technology. SciNet, a science-aware platform for research network operations, will curate multimodal network operations data and tag packets with science project information to allow correlation of data from heterogeneous sources, handle sensitive data securely and anticipate network requirements to deliver shorter experimental latency, fewer workflow failures and faster discovery cycles across science domains to satisfy workflow needs.
To learn more about this project, visit ornl.gov/projects/ai-powered-operations-self-driving-science-research










