A hybrid satellite architecture only has engineering value if the two bearers fail differently, scale differently, and serve distinct roles across the deployment lifecycle. Combining BGAN with 5G NR NTN exploits the specific propagation, protocol, and ecosystem characteristics of each path to maintain operational continuity while validating and scaling a standards-based system. This article examines what makes these two bearers complementary at the link and protocol level, and how that complementarity translates into concrete deployment architectures. For a broader perspective on why hybrid BGAN and NR NTN is gaining traction as a deployment strategy, the overview article covers the operational and programme-level rationale.
The engineering basis for complementarity
BGAN operates over Inmarsat GEO satellites at L-band, providing global coverage between approximately ±82° latitude. One-way propagation delay to geostationary orbit (35,786 km) is approximately 270–280 ms, yielding a radio-path RTT of roughly 540–560 ms before processing and core network traversal. Doppler on a GEO link is negligible: the satellite is stationary relative to the user, beam geometry changes slowly, and the channel does not exhibit rapid fading from satellite motion. BGAN failure modes are correspondingly predictable: terminal mispointing, L-band interference, or capacity exhaustion within a spot beam. These are understood, documented, and operationally manageable.
NR NTN, as defined in 3GPP Release 17, introduces a standards-based air interface designed for satellite access in both GEO and LEO configurations. At LEO altitudes (600–1,500 km), one-way propagation delay drops to 4–10 ms, but Doppler becomes significant: on the order of 1–4 kHz at 2 GHz carrier frequency, with variation rates up to tens of Hz/s depending on orbital geometry and elevation angle. The UE must perform Doppler pre-compensation and timing advance pre-computation based on GNSS position and satellite ephemeris broadcast in SIB extensions. The link is dynamic: beam footprints move, handovers are frequent, and channel conditions change on timescales of seconds to minutes. Even in GEO NR NTN configurations, the protocol stack differs fundamentally from BGAN. HARQ, RLC acknowledged mode, and RRC procedures must all accommodate RTT values that exceed terrestrial assumptions by one to two orders of magnitude.
The complementarity is structural. BGAN provides a stable, predictable bearer with known failure modes and minimal protocol-level complexity at the terminal. NR NTN provides a standards-aligned bearer with ecosystem integration, flexible deployment across orbit types, and a path toward capacity growth, but with protocol behaviours that are still being validated in operational environments.
How NTN-specific characteristics define the hybrid roles
The timing advance mechanism in NR NTN illustrates why the two bearers serve different roles. In terrestrial NR, timing advance compensates for propagation delay to ensure uplink signals arrive at the gNB within the cyclic prefix window, across cell radii of tens of kilometres. In NR NTN, the UE must pre-compute propagation delay using its own GNSS position and satellite ephemeris data received via SIB1 extensions (TS 38.331). For GEO NR NTN, this pre-compensation spans hundreds of milliseconds. For LEO, the value changes continuously as the satellite transits.
Doppler pre-compensation follows the same pattern. The UE computes the frequency offset from satellite velocity relative to the terminal and pre-corrects its transmit frequency so that the signal arrives at the gNB within the receiver’s tracking bandwidth, a procedure defined in TS 38.211 and TS 38.213. This depends on an accurate GNSS fix, current ephemeris, and correct geometry computation. Any error in pre-compensation, whether from stale ephemeris, degraded GNSS, or computational edge cases, produces a signal outside the gNB’s tolerance window.
BGAN has none of this complexity at the terminal. Its GEO link is essentially static in both timing and frequency. The terminal does not need GNSS for link maintenance, does not pre-compensate Doppler, and does not recompute timing advance dynamically. That stability is what makes BGAN a reliable continuity bearer while NR NTN procedures are validated under real-world conditions.
At the MAC layer, HARQ in NR NTN over GEO cannot function as a tight retransmission loop. With RTT exceeding 500 ms, the feedback cycle is too slow for HARQ to drive throughput recovery in the way it does terrestrially, where the HARQ feedback cycle is approximately 8–12 ms in sub-6 GHz terrestrial NR. Practical GEO NR NTN deployments reduce reliance on HARQ and shift retransmission responsibility to RLC acknowledged mode, with correspondingly extended t-PollRetransmit and t-Reordering timers. LEO NR NTN can sustain tighter HARQ operation, but the dynamic Doppler and timing environment introduces different edge cases. In both configurations, protocol behaviour under stress is less mature and less operationally characterised than BGAN’s simpler retransmission model.
Practical architecture: traffic steering, failover, and session continuity
A hybrid BGAN and NR NTN architecture requires explicit traffic classification by latency tolerance and reliability requirement. The engineering logic is direct.
Mission-critical signalling and low-bandwidth control traffic with strict continuity requirements routes over BGAN. The bearer’s stability, known failure modes, and predictable RTT make it suitable for flows where availability matters more than throughput. Bulk data, standards-validation traffic, and flows that benefit from ecosystem integration route over NR NTN, including traffic that exercises the 3GPP path for operational validation: testing HARQ behaviour under load, observing handover performance in LEO configurations, and characterising Doppler pre-compensation accuracy across elevation angles.
Failover trigger design must account for asymmetric path characteristics. A simple link-down trigger is insufficient. On the NR NTN side, degradation may manifest as increased residual Doppler error, timing advance drift, HARQ failure rate exceeding a threshold, or RRC re-establishment frequency. On the BGAN side, degradation is more likely to appear as throughput reduction from beam congestion or terminal pointing error. The failover logic must map these bearer-specific indicators to switching decisions rather than treating both paths as equivalent.
3GPP Release 16 introduced ATSSS (Access Traffic Steering, Switching and Splitting) as a framework for multi-access PDU sessions. Under ATSSS, a single PDU session anchored at the UPF can be served by multiple access legs. BGAN can be integrated as a non-3GPP access via an N3IWF or equivalent interworking function, while NR NTN provides native 3GPP access. ATSSS rules, provisioned via the PCF and enforced at the UE and UPF, enable flow-level steering based on 5QI, DNN, application identifier, or measured access quality, providing formal policy control over which flows use which bearer and under what conditions switching occurs.
Session continuity during bearer switchover depends on the ATSSS mode in use. In switching mode, active flows move from one access to another when the steering condition triggers. The IP anchor at the UPF remains unchanged, so upper-layer sessions survive. Interruption duration is bounded by how quickly the target bearer absorbs the flow, which for a pre-established BGAN session is determined by BGAN allocation delay rather than full session setup.
For staged migration, the hybrid architecture provides a defined path: begin with BGAN carrying operational traffic and NR NTN carrying validation traffic. As NR NTN behaviour is characterised and confidence builds, shift traffic classes progressively. Cutover criteria are defined by measured NR NTN performance against specific KPIs: HARQ success rate, handover completion rate, Doppler pre-compensation accuracy, and session continuity under mobility. The BGAN bearer remains available as a fallback throughout, with reversion triggered by threshold breach on any defined KPI.
Gatehouse perspective
Gatehouse has validated NR NTN protocol behaviour across both GEO and LEO delay and Doppler configurations. That work, exercising timing advance pre-computation, Doppler pre-compensation accuracy, HARQ operation under extended RTT, and RLC timer behaviour in realistic channel conditions, directly informs how the NR NTN side of a hybrid architecture should be tested and integrated.
The engineering challenge in hybrid deployments is understanding how NR NTN behaves at the edges of its operating envelope: where Doppler pre-compensation accuracy degrades at low elevation angles, where HARQ timers interact with bursty traffic under high RTT, where handover between beams produces transient loss. These behaviours determine when traffic can safely move from BGAN to NR NTN, and under what conditions it should move back. Test environments that reproduce real NTN link dynamics are necessary to characterise these boundaries.
The hybrid architecture’s value is the ability to maintain service continuity on a proven bearer while building operational evidence for a standards-based path, with defined criteria for when each bearer carries which traffic and validated protocol behaviour underpinning the transition.
Conclusion
BGAN and NR NTN are complementary because they occupy different positions across every axis that matters for deployment resilience: stability versus dynamism, known behaviour versus ecosystem alignment, operational maturity versus capacity growth. The hybrid value emerges from assigning each bearer a role matched to its characteristics, whether continuity assurance, validation, traffic steering, or staged migration, and defining the engineering criteria that govern transitions between them. The architecture is only as strong as the understanding of how each bearer behaves under stress, and how those behaviours interact when traffic moves between paths.
