[論文レビュー] Service Function Chain Routing in LEO Networks Using Shortest-Path Delay Statistical Stability

Li Zeng, Zixin Wang+2 more|arXiv (Cornell University)|Mar 4, 2026

Satellite Communication Systems被引用数 0

ひとこと要約

tldr: The paper introduces SA-MSGR, a stability-aware SFC routing method for LEO networks that uses pre-computed average shortest-path delays within a multi-stage graph to achieve lower and more predictable end-to-end SFC delays, outperforming baselines.

ABSTRACT

Low Earth orbit (LEO) satellite constellations have become a critical enabler for global coverage, utilizing numerous satellites orbiting Earth at high speeds. By decomposing complex network services into lightweight service functions, network function virtualization (NFV) transforms global network services into diverse service function chains (SFCs), coordinated by resource-constrained LEOs. However, the dynamic topology of satellite networks, marked by highly variable inter-satellite link delays, poses significant challenges for designing efficient routing strategies that ensure reliable and low-latency communication. Many existing routing methods suffer from poor scalability and degraded performance, limiting their practical implementation. To address these challenges, this paper proposes a novel SFC routing approach that leverages the statistical properties of network link states to mitigate instability caused by instantaneous modeling in dynamic satellite networks. Through comprehensive simulations on end-to-end shortest-path propagation delays in LEO networks, we identify and validate the statistical stability of multi-hop routes. Building on this insight, we introduce the Stability-Aware Multi-Stage Graph Routing (SA-MSGR) algorithm, which incorporates pre-computed average delays into a multi-stage graph optimization framework. Extensive simulations demonstrate the superior performance of SA-MSGR, achieving significantly lower and more predictable end-to-end SFC delays compared to representative baseline strategies.

研究の動機と目的

• Motivate NFV-based service delivery in dynamic LEO constellations with limited onboard compute.
• Investigate statistical stability of shortest-path delays in LEO networks.
• Develop a scalable routing algorithm that leverages delay stability for SFCs.
• Evaluate SA-MSGR against baseline strategies to demonstrate improved delay and predictability.

提案手法

• Demonstrate statistical stability of pairwise and multi-hop shorted-path delays via simulations of Walker Delta LEO constellations.
• Propose SA-MSGR that uses offline pre-computed average delays 0(D^tx) and a multi-stage graph to route SFCs.
• Construct a multi-stage graph with stages corresponding to VNFs and use pre-computed averages as edge weights.
• Compute the optimal SFC route by finding the shortest path on the DAG using dynamic programming.
• Analyze online complexity as O(M d7 S_max^2) per request, avoiding time-expanded graph scalability issues.

実験結果

リサーチクエスチョン

• RQ1 Do pairwise shortest-path delays D^tx(u,v,t) exhibit statistical stability in LEO networks (low coefficient of variation)?
• RQ2 Does the stability improve for longer multi-hop SFC paths (higher M)?
• RQ3 Can a stability-aware routing approach using pre-computed averages match or surpass time-expanded or snapshot-based methods in end-to-end SFC delay?
• RQ4 Is SA-MSGR computationally scalable for realistic SFC lengths and VNFs?

主な発見

• A large fraction of satellite pairs show low relative delay fluctuations; 70% have CV ≤ 0.2 and 90% have CV ≤ 0.3.
• Average path delay CV decreases with SFC length M, from 0.099 (M=1) to 0.028 (M=20).
• SA-MSGR consistently achieves the lowest average end-to-end SFC delay across tested M values, closely approaching the theoretical TEG-based method.
• SA-MSGR delivers highly stable delay distributions with narrow IQR and shorter whiskers compared to snapshot-based methods.
• Greedy-Cp generally outperforms Greedy-Tx, while Random performs worst; SA-MSGR and Snapshot-based methods outperform these heuristics overall.
• Online routing complexity is reduced by shifting optimization offline and using a DAG shortest-path on the MSG.