The Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks�ESnet Confab 2025�April 7-9, 2025

Julio Ibarra

AmLight Principal Investigator

Co-chair Vera Rubin Network Engineering Team

​

​

​

​

Outline

• Vera Rubin Observatory Service Level Agreement (SLA) and Operations Use Case
• Vera Rubin Observatory Long-Haul Network (LHN) Implementation
• Results from Image transfers
• Conclusion

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

2

Vera Rubin Observatory Long-Haul Network SLA

• The Vera Rubin Observatory Long-Haul Network (LHN) is an SLA-driven network, purpose-built to support the Vera Rubin network use case
• The LHN is built upon Research & Education (R&E) networks, collaborating to support the Vera Rubin SLA, and collectively referred to as the LHN Network Operators

• The LHN is to provide a guaranteed 40Gbps end-to-end network transport service from the Summit to the US Data Facility at SLAC
• End-to-end service availability of 99%
• Mean Time to Restore Service (MTTR) is not to exceed 4 hours
• Network measurement instrumentation
• Troubleshooting, Active measurements, network telemetry
• Export data to the Rubin Virtual NOC (VNOC)
• Participation in the Rubin NET team

​

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

3

Use Case: Vera Rubin Observatory Operation��

​

• The camera will take a picture of the southern sky every 27 seconds, and produce a 13GB image

​

• Each image must be transferred to the USDF at SLAC, within 7 seconds, inside the 27 second transfer window

​

• Constraints
• Distance from the Base station to the USDF is approximately 12,000 miles
• RTT from the Base Station to the USDF is approximately 180+ ms
• 0.001% of packet loss will compromise the Rubin Observatory image transfer workflow

​

• Challenges
• How to build the LHN to deliver a 13GB image within the 7s window?
• How to recover from an event that impacts network service on the LHN, and continue to deliver 13GB images on time?

​

​

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

4

Vera Rubin LHN: High-level Topology

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

5

LHN Network Operators:

Vera Rubin LHN: Detailed View

Results: Transfers over time���Roadmap to Operations�

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

7

Results by Wil O’Mullane

Vera Rubin Observatory

Min transfer time 4s

Max transfer time 51s

7s desired transfer time

51s max transfer time

34s absolute time before images crash

Images arriving within 4s and 15s

34s absolute time before images bump into each other

Tail of transfers to reduce within 7s

Conclusion

• The Vera Rubin LHN is operating and complying with the SLA requirements
• Testing, Testing! NSF–DOE Vera C. Rubin Observatory Completes Comprehensive System Tests With Flying Colors

​

• The LHN is highly instrumented with PerfSonar nodes, packet and optical telemetry, and protocols for monitoring network continuity and end-to-end control to
• Detect potential network service interruption, and to
• Respond before the image transfer window is compromised

​

• The Vera Rubin Virtual NOC (VNOC) is providing NOC services to the LHN Network Operators
• The VNOC is working towards providing NOC services to the Vera Rubin Observatory

​

ESnet Confab 2025: Vera Rubin Observatory Long-Haul Network: Powered by Research and Education Networks

8

THANK YOU

More Slides Follow

Vera Rubin LHN: Logical Network Topology

Results: Ingest over time

Innovating the Network for Data Intensive Science 2019

12

4.8s Min Ingest Time

34s absolute time before images crash

Median time between shutter close and ingest: 11s

Mean time between shutter close and ingest: 12s

Results by Wil O’Mullane

ESnet Confab 2025

13