Markdown Version | Transcript | Session Recording | Session Materials
Session Date/Time: 17 Mar 2026 08:00
SPACE
IETF 125 Session: Systems and Protocol Aspects for Circumstellar Environments (Proposed Research Group)
Summary
The SPACE proposed Research Group (RG) met to discuss long-term research regarding space-based networking, orbital computing, and interstellar protocol architectures. The session covered the integration of terrestrial cloud infrastructures with space segments, standardized protocol stacks for Earth-Moon communication, and tools for LEO measurement. Technical presentations also introduced a standardized code for describing satellite constellations and novel approaches for sparse LEO networking and space-based AI infrastructure.
Key Discussion Points
Space Cloud Network Infrastructure
Kanglian Zhao presented Bridging the Future Digital Divide with Space Cloud Network Infrastructure.
- Infrastructure Evolution: Modern space initiatives (Starlink, Amazon Kuiper, Blue Origin's TerraWave) are shifting from "bent-pipe" architectures to integrated space-terrestrial clouds with orbital data centers.
- Use Cases: Identified needs for distributed AI model training in orbit (e.g., EU ASCEND project) and space-based Point of Presence (POP) to reduce latency and ensure data sovereignty.
- Architecture: Proposed a two-layer model consisting of a Cloud Layer (for computation/storage) and an Access Layer (for connectivity). Discussion focused on orbit selection, specifically the benefits of Dawn-Dusk orbits for consistent solar power.
Earth-Moon Integrated Network Protocols
He Xiongwen presented Protocol Architecture of Earth-Moon Integrated Network.
- Heterogeneous Networking: Deep space missions (Chang'e, Tianwen) require bridging CCSDS, ECSS, and IETF protocol suites.
- Proposed 4-Layer Architecture:
- Application Layer: Support for CFDP and ECSS-PUS.
- Transport Layer: Support for TCP, UDP, and LTP.
- Network Layer: Integration of Space Packet Protocol (for low-speed links) and IPv6 (for high-speed optical/TSN links).
- Link/Physical Layer: Convergence of diverse space-to-ground and inter-satellite links.
- Implementation: Mentioned the "Fuxi" software architecture and the deployment of Intelligent Network Units on the Queqiao-2 relay satellite.
Tools and Measurement
Nishanth Sastry presented Tools and Infrastructures for LEO Measurement and Evaluation.
- LEOscope: A testbed using Docker-encapsulated experiments deployed behind satellite user terminals.
- Observation: Measuring mega-constellations requires global vantage points; researchers currently rely on browser plugins, public datasets (M-Lab, RIPE Atlas), and simulators.
- Community Call: Requested contributions to a living tools document. Jörg Ott suggested starting with a wiki for rapid population before moving to an Internet-Draft.
Standardizing Constellation Descriptions
Juan Fraire presented Draft update: A Code to Describe Satellite Constellations.
- Draft: draft-fraire-space-constellation-code-01
- Grammar: Uses ABNF to describe Walker Star and Walker Delta topologies (altitude, inclination, planes, phasing).
- Connectivity Patterns: Introduced a YAML-based extension to define link patterns (e.g., checkered grid patterns) using simple predicates.
- Future Work: Considering "Flower" constellations (eccentric orbits) and optical terminal constraints (gimbal/slew limits).
IP-Based Satellite Routing and Computing
Jing Wang presented two topics:
- Consideration for IP-Based Satellite Routing Protocol (draft-wang-space-satellite-routing-considerations-00).
- Challenges include dynamic topologies and limited spaceborne resources.
- References potential integration with TVR (Time-Variant Routing) and LISP (Locator/ID Separation Protocol).
- Consideration for Space-Based Computing Infrastructure Network (draft-wang-space-computing-infrastructure-00).
- Proposed requirements for monitoring onboard resources and computing-aware routing to prevent node overload during high-demand events (e.g., disaster monitoring).
Small-Scale LEO Scaling
Yuanjie Li presented Small-scale LEO Satellite Networking for Global-scale Demands.
- Problem: High cost and orbital debris risks associated with mega-constellations.
- Tiny LEO: Proposed using "Compressed Sensing" algorithms to design non-uniform, sparse constellations that match global demand density while using significantly fewer satellites.
- Geographic Segment Anycast: To handle frequent topology changes, a routing abstraction was proposed using SRv6 to decouple high-level geographic intent from low-level satellite mobility.
- Safety: Discussed "cascade collision avoidance" where maneuvers in a dense constellation trigger a domino effect of safety adjustments, potentially exhausting fuel budgets.
Decisions and Action Items
- Tools Collection: The group agreed to establish a centralized repository/wiki for tools and datasets relevant to space networking research.
- Constellation Code: Contributors are encouraged to review draft-fraire-space-constellation-code to ensure the YAML structure is expressive enough for varied link constraints (e.g., space-to-ground).
Next Steps
- Draft Progression: Authors of the routing and computing drafts (draft-wang-*`) and the constellation code draft will incorporate feedback from the session and the mailing list.
- IETF 126 Preparation: The chairs (Juan Fraire and colleagues) will look to prioritize topics for the next meeting in Vienna, focusing on converging the various tooling and description format efforts.
Related Documents
draft-fraire-space-constellation-code, draft-fraire-space-constellation-code-01, draft-update-a-code-to-describe-satellite-constellations-04, draft-wang, draft-wang-space-computing-infrastructure-00, draft-wang-space-satellite-routing-considerations-00