Network Flow and Packet Marking for Global Scientific Computing
It provides an open system using open source technologies that helps Research and Education (R&E) providers in understanding how their networks are being utilised while at the same time providing feedback to the scientific community on what network flows and patterns are critical for their computing.
Our approach is based on a network tagging mechanism that marks network packets and/or network flows using the science domain and activity fields. These tags can then be captured by the R&E providers and correlated with their existing netflow data to better understand existing network patterns, estimate network usage and track activities.
The initiative offers an open collaboration on the research and development of the packet and flow marking prototypes and works in close collaboration with the scientific storage and transfer providers to enable the marking capability. The project is currently in the prototyping phase and is open for participation from any science domain that require or anticipate to require high throughput computing as well as any interested R&E providers.