Multilayered Satellite Networks

• Multilayered Satellite Networks are integrated architectures combining spaceborne and aerial platforms (GEO, MEO, LEO, HAP, UAV) to provide persistent, flexible connectivity.
• They employ advanced inter-layer links using technologies such as FSO/RF, SDN, and NFV to optimize latency, throughput, and network resilience.
• MLSNs enable scalable applications like broadband access, disaster response, and IoT connectivity through efficient network slicing and dynamic resource management.

Multilayered Satellite Networks (MLSN) are spatially and functionally integrated communication architectures comprising multiple tiers of spaceborne and aerial platforms, typically including Geostationary Earth Orbit (GEO), Medium Earth Orbit (MEO), Low Earth Orbit (LEO) satellites, High Altitude Platforms (HAPs), Uncrewed Aerial Vehicles (UAVs), and terrestrial nodes. MLSNs leverage heterogeneous orbits, dynamic interconnections, advanced backhaul/fronthaul technologies (e.g., Free-Space Optical (FSO) and Radio Frequency (RF)), and programmable networking to deliver high-throughput, resilient, and low-latency connectivity over wide geographical areas. The inherently multilayer design enables persistent coverage, rapid-reaction failover, flexible Quality of Service (QoS) provisioning, and efficient integration with terrestrial and non-terrestrial networks.

1. Multilayered Satellite Network Architectures

MLSNs can be systematically decomposed into vertical layers, each characterized by orbital altitude, coverage footprint, platform type, and functional role. Dominant architectural patterns include concentric-shell configurations (LEO/MEO/GEO) and hybrid multilayer constructs incorporating HAPs and UAVs.

• LEO layer: Dense constellations (e.g., Starlink, Telesat) at 500–2,000 km. Characteristics: low latency (5–50 ms), small coverage, high Doppler, frequent handovers, direct-to-ground and inter-satellite link (ISL) capability (Hu, 2022, Hu, 2023, Al-Hraishawi et al., 2021).
• MEO layer: Regional constellations (e.g., O3b) at 2,000–10,000 km. Provides bridging between LEO and GEO, stability for inter-layer links, and reduced handover rates (Hu, 2022, Wang et al., 2020).
• GEO layer: At ~35,786 km, broad “umbrella” coverage, continuous visibility, highest latency (~250 ms), backbone for control and high-throughput backhaul (Hu, 2023, Wang et al., 2020).
• HAP/UAV sub-layers: Stratospheric HAPs (~20 km) and UAVs (sub-km to several km) provide regional relaying, edge coverage, local traffic offloading, and last-mile reach (Li et al., 17 May 2025, Khennoufa et al., 22 Feb 2025, Wang et al., 2020).

Inter-layer connectivity is established through dynamic or static cross-layer links (FSO, RF, millimeter wave), with topological models often formalized as supra-graphs linking layer-specific subgraphs (Hao et al., 2023).

2. Intra- and Inter-Layer Connectivity: ISLs and ILCs

• Intra-layer connectivity is realized by ISLs, forming ring, mesh, or grid topologies within each orbital shell:
  • LEO: Typically, four ISLs per satellite (two intra-plane, two inter-plane) for polar or Walker–Delta constellations (Hu, 2022, Al-Hraishawi et al., 2021).
  • MEO and GEO: Full mesh and/or backbone rings (Hu, 2022).
• Inter-layer connectivity (ILCs) creates vertical coupling between adjacent shells (e.g., LEO–MEO, LEO–GEO, HAP–LEO).
  • ILC deployment is constrained by visibility, relative motion (“co-motion”), and resource limits (e.g., maximum number of transceivers per satellite) (Hao et al., 2023).
  • The optimal number of ILCs per adjacent pair is proven to be below half the smaller layer’s size, with uniform deployment on orbital planes minimizing average path length and handover events (Hao et al., 2023).
• Two-phase ILC deployment (symmetry reduction and handover-aware prioritization) yields polynomial-time selection algorithms achieving near-optimal hop reduction and throughput gains of up to 1.55× versus isolated layers (Hao et al., 2023).

Table 1. Typical Layer Parameters and Roles

Layer	Altitude (km)	Role	Connectivity Type
GEO	~36,000	Backhaul/Control	GEO–GEO ISLs, ILCs
MEO	2,000–10,000	Regional Relay/Bridge	MEO–MEO ISLs, ILCs
LEO	500–2,000	Access, Edge, Backhaul	LEO–LEO ISLs, ILCs
HAP	~20	Regional Edge/Relay	HAP–LEO, HAP–UAV
UAV	0.1–2	Last-mile/Edge	UAV–HAP, UAV–UE

3. Networking Protocols, Access Schemes, and Software Control

• Access and multiple access: Orthogonal schemes (TDMA, FDMA, CDMA), spatial division (SDMA, UM-MIMO), and advanced non-orthogonal multiple access (NOMA, RSMA). RSMA, deployed in distributed mode (D-RSMA), manages inter-layer and intra-layer interference robustly, supporting max-min fairness via rate splitting among super-common, sub-common, and private messages (Xu et al., 2023).
• Channel models: Layer-specific path loss (FSPL, shadowed-Rician, Nakagami-m), link-specific attenuation (gas, rain, scintillation), and hardware impairment models (EVM, phase noise, I/Q imbalance). RIS-aided links introduce further beamforming gains (∼ρ²N²) and specific double path-loss effects (Khennoufa et al., 22 Feb 2025).
• Programmable control: SDN and NFV-based orchestration (hierarchical controller distribution: GEO as global, MEO as domain, LEO as local), supporting per-service network slicing, agile rerouting, dynamic resource allocation, and seamless handovers (Al-Hraishawi et al., 2021).
• Backhaul and routing: Multi-commodity flow ILPs, with objectives that minimize cost, delay, or hop count under capacity and latency constraints, are common. Software-defined strategies support rapid snapshot-based path reconfiguration during topology changes (Al-Hraishawi et al., 2021, Hao et al., 2023).

4. Performance Analysis: Latency, Reliability, Throughput, and Fairness

• Latency: Multi-hop, multi-layer routing dramatically reduces end-to-end latency versus direct ground–satellite contact. LEO–LEO–LEO paths achieve sub-10 ms for short packets; GEO–LEO or GEO–LEO–GEO paths generally provide sub-100 ms, depending on payload/data rate (Hu, 2023, Hu, 2022).
• Reliability and resilience: End-to-end reliability $\Phi$ increases with parallel/hybrid path utilization across layers (e.g., LEO–LEO + GEO–LEO), achieving >99.99% under optimal assignment, compensating for lower per-hop LEO reliability ($\phi_{LEO–LEO}\sim0.998$). Multi-layer failover delivers 100% mission resilience compared to <8% for direct access in low-visibility constellations (Hu, 2022).
• Throughput: Spatial cooperation via well-placed ILCs yields up to 1.55× aggregate throughput over isolated layer operation. Average hop-count decays logarithmically with ILC count, subject to physical and visibility constraints (Hao et al., 2023).
• User fairness: D-RSMA schemes in GEO/LEO networks outperform SDMA and NOMA by 7–14% in max-min fairness for moderate terminal density, with larger gains in dense LEO deployments and under imperfect CSI (Xu et al., 2023).
• Energy efficiency and coverage: RIS-aided UAV/HAP deployments elevate coverage probabilities (>0.9) and boost aggregate throughput (up to 200 Mbps/cell) for rural and emergency scenarios, though hardware impairments can reduce effective gains (Khennoufa et al., 22 Feb 2025).

5. Optimization and Learning-Based Management

• Joint resource optimization: Typical problems scalarize objectives such as maximizing downlink throughput, minimizing handover events, and balancing load under path constraints and dynamic topology (Li et al., 17 May 2025, Hao et al., 2023).
• DRL-based control: LLM-guided Truncated Quantile Critics with Dynamic Action Masking (LTQC-DAM) adaptively optimizes satellite handover and resource allocation, outperforming standard Deep RL in convergence speed, handovers (17.7% reduction), and final throughput (up to 0.44% gain) in multi-tier LEO–HAP–ground networks with hybrid FSO/RF links (Li et al., 17 May 2025).
• Hyperparameter auto-tuning: Utilizing LLM meta-functions (e.g., DeepSeek) for context-aware online parameter adaptation stabilizes RL learning, reduces exploration inefficiency, and avoids late-stage oscillations observed in conventional techniques (Li et al., 17 May 2025).
• Interference/handovers: Advanced cross-layer masking and prompt-based LLM-driven tuning enable exploration focused exclusively on valid, visible satellites, which is critical in dynamic, partial-observation environments (Li et al., 17 May 2025).

6. Use Cases and Practical Deployments

• Broadband connectivity: Rural broadband via HAP–UAV–ground with NOMA achieves coverage radii exceeding 100 km, aggregate cell throughput of 100–200 Mbps, and low-latency (<50 ms) access (Khennoufa et al., 22 Feb 2025).
• Disaster response: On-demand UAV-TRIS clusters under HAPs serve as emergency backbones, delivering >50 Mbps and sub-30 ms critical message latency, independent of terrestrial infrastructure (Khennoufa et al., 22 Feb 2025, Li et al., 17 May 2025).
• Telemetry and telecommand (TM/TC): Batch and real-time TM/TC monitored over O(24 hr) scenarios demonstrate sub-15 ms average latency, 99.2% reliability, and full resilience using three-layer routing and hybrid LEO–MEO–GEO fallback; operating costs are reduced via minimization of ground station contacts (Hu, 2022, Hu, 2023).
• Nanosatellite and IoT access: MLSNs with LEO/MEO/GEO and programmable SDN/NFV infrastructure deliver continuous broadband to mobile nanosat users with >80% persistent coverage and rationalized SDN control-plane overhead (Al-Hraishawi et al., 2021).

7. Open Challenges and Future Directions

• Mobility and handover management: Seamless intra/inter-layer handovers remain a challenge due to rapid topology evolution; controller load and protocol overhead in high-frequency LEO constellations require further research (Hao et al., 2023, Al-Hraishawi et al., 2021).
• Inter-layer coordination: Dynamic selection of relaying paths and cross-layer scheduling is an unsolved NP-hard problem, especially for multi-objective designs balancing delay, reliability, and energy (Hao et al., 2023, Wang et al., 2020).
• Hardware and channel impairments: Robust interference management, compensation for phase-noise, IQ imbalance, and FSO/rain fading are key for realistic operations (Khennoufa et al., 22 Feb 2025, Xu et al., 2023).
• Distributed intelligence: Next-generation solutions are likely to integrate RL/LLM-powered optimization, AI-driven impairment tracking, and continual network slicing for real-time, resource-efficient orchestration (Li et al., 17 May 2025, Al-Hraishawi et al., 2021).
• Standardization and sustainability: Coordination with 3GPP NTN Release 18, SDG alignment (carbon footprint, recyclability), and new holistic protocol stacks for space–air–ground integration remain open (Khennoufa et al., 22 Feb 2025).

• (Hu, 2022) Hu et al., "Enabling Resilient and Real-Time Network Operations in Space: A Novel Multi-Layer Satellite Networking Scheme"
• (Hu, 2023) Zhang et al., "Toward Multi-Layer Networking for Satellite Network Operations"
• (Hao et al., 2023) Ju et al., "High Throughput Inter-Layer Connecting Strategy for Multi-Layer Ultra-Dense Satellite Networks"
• (Xu et al., 2023) Liu et al., "Distributed Rate-Splitting Multiple Access for Multilayer Satellite Communications"
• (Li et al., 17 May 2025) Guo et al., "LLM-guided DRL for Multi-tier LEO Satellite Networks with Hybrid FSO/RF Links"
• (Khennoufa et al., 22 Feb 2025) Jian et al., "A Multi-layer Non-Terrestrial Networks Architecture for 6G and Beyond under Realistic Conditions and with Practical Limitations"
• (Al-Hraishawi et al., 2021) Elmegyawi et al., "Multi-layer Space Information Networks: Access Design and Softwarization"
• (Wang et al., 2020) Chien et al., "The Potential of Multilayered Hierarchical Nonterrestrial Networks for 6G: A Comparative Analysis Among Networking Architectures"