Non-Terrestrial Network (NTN) and Satellite Communication
1. Overview
A Non-Terrestrial Network (NTN) refers to a network that uses airborne or space-based payloads rather than terrestrial base stations—satellites (LEO·MEO·GEO), High-Altitude Platform Stations (HAPS), unmanned aerial vehicles, and so on—as radio access nodes to deliver cellular connectivity to devices within the 3GPP 5G/6G standard framework. The essence is integrating satellites into the same standards, core, and device ecosystem as the Terrestrial Network (TN), rather than treating them as a separate closed system.
The reason NTN rose onto the standards stage is the geographic coverage gap that terrestrial cellular networks structurally cannot overcome. Base stations are built only where power, backhaul, land, and profitability are secured, so areas such as oceans, deserts, mountains, polar regions, and flight routes remain radio shadows even in the 5G era. Statistically, more than 70% of the Earth's surface is sea and much of the land is sparsely populated, so "area-based coverage" paints an entirely different picture from "population-based coverage." Even if a terrestrial network covers 95% of the population, more than half of a nation's land area may receive no signal, and the demand to fill precisely this gap is NTN's primary driver.
The necessity becomes clear in three directions. First is ubiquitous connectivity, guaranteeing anywhere the minimal communication that is a matter of life and death in disasters, distress, and remote areas. Since smartphone makers began embedding "satellite emergency messaging" features after 2022, the benefit of being able to send a rescue request even when terrestrial networks are fully down has been imprinted on the public. Second is resilience: when terrestrial base stations are destroyed or backhaul is severed by earthquakes, floods, or power outages, satellites provide an alternative and complementary path. Third is industrial demand: demand is exploding in high-mobility, infrastructure-less environments—shipping, aviation, energy, defense, and IoT sensor networks—that need wide-area low-rate connectivity or broadband backhaul.
The essential characteristics that define NTN are long propagation delay, fast-moving cells, and standard device compatibility. A GEO satellite has a round-trip propagation delay of hundreds of milliseconds, reshaping the design of latency-sensitive applications, while an LEO satellite moves at 7–8 km/s, so the cell itself constantly moves, forcing frequent handovers and Doppler compensation. Even so, the most important aim of 3GPP NTN is to let an unmodified handset connect directly, and this "standards integration" is the decisive dividing line that separates NTN from the earlier dedicated satellite phones (Iridium, Inmarsat).
2. Overall Structure and Components of NTN
The key to understanding NTN lies in the structural change of inserting a satellite relay segment between the base-station-to-core path of the terrestrial network. The radio segment linking the device (UE) and the satellite is called the service link, the segment linking the satellite and the ground gateway is called the feeder link, and the gateway connects this to the core network (5GC). The overall structure diagram below shows this three-stage composition in plain terms.
flowchart TB
UE["UE device(smartphone·IoT)"] -- "service link(radio)" --> SAT["satellite payload(LEO/MEO/GEO)"]
SAT -- "feeder link(radio)" --> GW["ground gateway"]
GW -- "wired backhaul" --> CORE["5G core(5GC)"]
CORE --> DN["data network·internet"]
SAT -. "inter-satellite link(ISL)" .-> SAT2["adjacent satellite"]
SAT2 -- "feeder link" --> GW2["remote gateway"]
Examining the components one by one reveals why each exists. The device (UE) is the starting point that determines NTN's success or failure; the link budget must be designed so that even devices with small transmit power and antenna gain, like ordinary phones, can connect. Because the distance to the satellite ranges from hundreds to tens of thousands of km, the path loss is overwhelmingly larger than in terrestrial networks, and to compensate, a large phased-array antenna is carried on the satellite to raise beam gain, while the device is offered limited services prioritizing low-rate, narrowband transmission (IoT-NTN) or emergency-message levels.
The satellite payload is the heart that relays the service and feeder links, and it divides into the transparent and regenerative methods discussed in detail later. The gateway is the junction between the satellite and the ground core, and since one satellite passes through the visibility range of several gateways, feeder link switchover between gateways occurs frequently. The Inter-Satellite Link (ISL) is a core element that lets satellites communicate directly by optical or radio means, so that even without a nearby gateway, traffic can be routed in space to a distant gateway. In practice, ISL is the key that makes coverage possible in areas with no ground gateway at all, such as the middle of an ocean.
Because the design trade-offs change completely with orbital altitude, orbital characteristics must be compared while also pointing out the reason the differences arise. LEO, being close, has small delay and path loss, favoring ordinary-device connection and low-latency applications, but due to its fast movement it must form a constellation of many satellites (hundreds to thousands) for continuous coverage, with frequent handovers. GEO is synchronized with Earth's rotation about 35,786 km above the equator, providing a fixed cell with almost no handover, but its round-trip delay reaches hundreds of ms, disadvantaging real-time performance and TCP.
| Category | LEO (low orbit) | MEO (medium orbit) | GEO (geostationary) |
|---|---|---|---|
| Altitude | approx. 300–2,000 km | approx. 8,000–20,000 km | approx. 35,786 km |
| One-way delay | a few ms (few–tens of ms) | tens of ms | approx. 120–140 ms (250 ms+ round trip) |
| Cell movement | very fast (7–8 km/s) | fast | fixed relative to ground |
| Satellites needed | many (constellation required) | moderate | few (3 for whole Earth) |
| Suitable services | broadband·low-latency·D2C | positioning·moderate delay | broadcast·IoT·wide-area backhaul |
The practical implication revealed by this comparison is that no single orbit can satisfy all requirements. Thus actual operators adopt a multi-orbit strategy combining LEO's low latency with GEO's wide-area stability, and a traffic-nature-based design is taking hold—distributing latency-sensitive voice and gaming to LEO, and delay-insensitive IoT and broadcast to GEO.
Meanwhile, NTN is not limited to satellites. High-Altitude Platform Stations (HAPS) that stay at about 20 km in the stratosphere and unmanned aerial vehicles (UAVs) are also defined as NTN payloads; being far closer to the ground than satellites, they have small path loss and delay but a narrow coverage radius and constraints on flight duration. HAPS can stay over a specific region for long periods and provide stable coverage like a ground base station, making them especially effective for localized, temporary demand such as a temporary base station in a disaster area or short-term capacity reinforcement at a large event venue. In this way, NTN should be understood as a set of airborne and space nodes layered by altitude, from orbital satellites to stratospheric platforms, and this altitude diversity leads to 6G's vision of a 3D integrated network.
3. Payload Architecture: Transparent vs Regenerative
The most important design decision governing NTN's performance and complexity is "whether to keep the base station's signal processing on the ground or carry it on the satellite." Depending on this choice, the satellite becomes either an "antenna in the sky" or a "base station in the sky." The detailed architecture diagram below contrasts the functional placement of the two methods.
flowchart LR
subgraph TR["transparent payload(bent-pipe)"]
U1["UE"] --> S1["satellite: frequency conversion·amplification only"]
S1 --> G1["gateway: gNB on the ground"]
G1 --> C1["5GC"]
end
subgraph RE["regenerative payload"]
U2["UE"] --> S2["satellite: gNB onboard(demod·remod·routing)"]
S2 --> G2["gateway: simple relay"]
G2 --> C2["5GC"]
end
The transparent payload is a "bent-pipe" method in which the satellite only performs frequency conversion and amplification and reflects the received signal back as is. All base station (gNB) signal processing remains at the ground gateway and the satellite acts purely as a repeater, so the satellite hardware is simple and less affected by standards changes, favoring early commercialization and extended satellite lifespan. 3GPP Release 17 NTN adopted this transparent method as its baseline. However, because the service link and feeder link are serially connected into a single radio path, both segments add to the total delay, and service is cut off once the gateway's visibility range is left.
The regenerative payload is a "base station in the sky" method that carries all or part of the base station (gNB) function on the satellite, demodulating and decoding received radio signals and then remodulating and routing them. Since the satellite regenerates the signal directly, the influence of feeder link quality on the service link is reduced, and combined with ISL it can complete a path in space even in areas without a gateway, approaching true global coverage. In return, a high-performance computing unit must be carried on the satellite, imposing heavy power, heat, and cost burdens, and software and hardware updates become difficult each time the standard is revised. 3GPP has been evolving in this direction, progressively standardizing the regenerative payload, store-and-forward, and full gNB onboard in Releases 18–19.
| Comparison item | Transparent payload | Regenerative payload |
|---|---|---|
| Satellite function | frequency conversion·amplification (repeater) | demod·remod·routing (gNB onboard) |
| Satellite complexity | low | high (onboard computation) |
| Delay characteristics | two links accumulate serially | optimizable via link separation |
| ISL utilization | limited | global via space routing |
| Standard | Rel-17 baseline | expanded from Rel-18/19 |
What the difference between the two methods means in practice is a trade-off between commercialization speed and coverage completeness. Early Direct-to-Cell services adopt a simple design for quick launch, but to expand coverage and capacity they must eventually move to a regenerative type with onboard processing and ISL. In fact, some leading operators' early direct-connection satellites started as regenerative types carrying LTE (eNodeB) modems and foreshadow an upgrade to 5G, showing that the "base station in the sky" is no longer a concept but a reality operating in orbit.
4. Core Technical Challenges of NTN and Solutions
NTN is the work of porting the 5G NR specification designed for terrestrial networks into an environment with extreme distance, mobility, and delay, so unique challenges absent in terrestrial networks arise. The following four are key issues that must be addressed in depth in an exam answer.
A. Large Propagation Delay and Timing Advance
In a terrestrial network the device-to-base-station distance is at most tens of km, so the round-trip delay is mere hundreds of microseconds, but in NTN even LEO has a round-trip of several ms and GEO reaches hundreds of ms. 5G NR's random access, HARQ, and timing advance procedures did not assume such large delays, so NTN changes the design to broadcast the satellite's position and ephemeris to the device so that the device pre-compensates its own delay to the satellite. Specifically, the device determines its own position via GNSS and computes the round-trip delay using the broadcast satellite orbital information to apply a large timing advance, while HARQ is relaxed by increasing the number of processes to accommodate the large round-trip time (RTT) or by disabling feedback.
B. Doppler Shift and Frequency Compensation
Because an LEO satellite moves at 7–8 km/s, as seen by the device the received frequency undergoes a Doppler shift that changes abruptly on the order of tens of kHz. This is of a different dimension from the Doppler experienced at the low speeds of terrestrial networks (tens to hundreds of km/h), and without compensation demodulation itself becomes impossible. NTN designs the device and the satellite to each pre-compensate for Doppler based on satellite orbital information, and especially in Direct-to-Cell where ordinary smartphones connect, the device cannot be modified, so the satellite-side SDR absorbs both Doppler pre-compensation and timing so that the device modem sees only a tolerable error range. This very design philosophy that "the satellite absorbs the complexity" is the secret to standard-device compatibility.
C. Frequent Handovers and Service Continuity
GEO has a fixed cell with almost no handover, but in LEO a single cell (beam) illuminates a given point for only a few minutes, so constant handovers occur. Even if the device stays still, the satellite passes by and the cell changes, so both satellite-to-satellite (NTN-NTN) and satellite-to-ground (NTN-TN) mobility must be managed. 3GPP strengthened conditional handover (CHO), orbital-information-based predictive handover, and the service continuity specification between NTN and terrestrial networks in Release 18, designing the device to move seamlessly from satellite to ground cell or vice versa. Also, in the regenerative type, even if the feeder link is cut, the satellite stores data via store-and-forward and delivers it to the next gateway, complementing the continuity of delay-tolerant services.
Unlike terrestrial handover, which is triggered "when the device moves and the signal weakens," NTN handover is essentially different in that it is prepared in advance at the predicted moment that "the satellite will pass by and soon leave the visibility range." The procedure diagram below shows the conditional handover flow in which the device continuously switches to the next satellite when the LEO satellite is about to set below the horizon.
sequenceDiagram
participant UE as UE device
participant S1 as satellite A(setting)
participant S2 as satellite B(rising)
participant GW as gateway·5GC
GW->>UE: broadcast ephemeris·CHO conditions(pre-configure satellite B)
UE->>UE: predict switch timing from own position·ephemeris
S1->>UE: service continues(signal begins to weaken)
UE->>S2: switch to satellite B when condition met
S2->>GW: path re-establishment(via feeder/ISL)
GW-->>UE: maintain service continuity(uninterrupted)
D. Link Budget and Coverage Enhancement
To let a small device overcome the enormous path loss to the satellite, the link budget design is critical. On the satellite side, beam gain is raised with a large phased-array antenna of hundreds to thousands of elements, and multiple beams are projected onto the ground to reuse frequency. On the device side, coverage enhancement techniques such as low-rate, narrowband transmission, repetition, and low modulation order lower the signal-to-noise ratio requirement. Releases 18–19 address uplink and downlink coverage enhancement and uplink capacity improvement as explicit tasks, advancing beyond the emergency-message level toward practical broadband.
5. Comparison and Application Cases
NTN's benefit emerges most sharply on two axes: "Direct-to-Cell/Device," which uses an unmodified ordinary device, and low-rate wide-area connectivity for IoT sensor networks (IoT-NTN). Both axes share the common purpose of "filling the terrestrial network gap," but their designs diverge because the required bandwidth and device characteristics differ.
Direct-to-Cell is a method in which the satellite directly transmits signals to an ordinary smartphone using the same frequency and standard as the ground, and the satellite acts like a "cell in the sky" without a special antenna or app on the device. Representatively, one LEO operator designed satellites carrying software-defined radios and large phased arrays to act like LTE/NR macro cells in the air, while another operator aims for multi-Mbps-class links with very large phased arrays. As of 2026, some constellations are reported to operate hundreds of direct-connection satellites in orbit, serving dozens of countries in a roaming form, but detailed figures and commercial scope vary by operator and point in time, so assertions should be avoided. Currently it starts with text and emergency messages and expands to voice and low-rate data, positioning itself as a complement that fills gaps rather than replacing terrestrial network quality.
IoT-NTN connects low-power wide-area (LPWAN) devices such as NB-IoT and eMTC via satellite, collecting sensor data from places terrestrial networks do not reach, such as ocean buoys, pipelines, agriculture, and logistics containers. With small required bandwidth and insensitivity to delay, it fits well with GEO or store-and-forward methods, and coverage enhancement techniques are especially important because device price and power budget are tight. In disasters, even if terrestrial communication is paralyzed, IoT-NTN sensors transmit river water-level and wildfire-detection data via satellite, providing the concrete benefit of raising disaster resilience.
Comparing NTN with earlier dedicated satellite phones (Iridium, Thuraya, etc.) and terrestrial 5G reveals the essence of the differences. Dedicated satellite phones were closed with proprietary specifications and dedicated devices, making them expensive and limited in adoption, whereas NTN uses 3GPP standards and ordinary devices, so it has the decisive advantages of economies of scale and ecosystem compatibility. Compared with terrestrial 5G, NTN is inferior in delay and capacity but overwhelmingly superior in coverage and resilience, so the two converge not into competition but into an integrated network (TN+NTN) in which satellites complement the terrestrial network, and this is the key point.
6. Deep Dive: 3GPP Standards Trends and 6G Integration
NTN's standardization flow is summarized by the evolution along two axes: "from complement to integration, from transparent to regenerative." 3GPP first standardized NTN in Release 17 (2022), introducing NR-NTN (handheld, broadband-oriented), which extends 5G NR, and IoT-NTN (NB-IoT, eMTC), which extends the 4G family, on a transparent payload basis. This specification was generalized to encompass LEO, MEO, and GEO, securing flexibility not bound to a specific orbit.
Release 18 (5G-Advanced) strengthened coverage enhancement, operation above 10 GHz (especially the Ka band), and mobility and service continuity between NTN and terrestrial networks (NTN-TN) and between satellites (NTN-NTN), and addressed store-and-forward based on a regenerative payload and feeder link switchover. Release 19 is the next phase of NR-NTN, including downlink coverage enhancement, uplink capacity and throughput optimization, NR-NTN broadcast (MBS), and support for a regenerative payload carrying a full gNB, with a freeze expected around 2025 (detailed scope and schedule may change as the specification is not yet finalized).
The terminus of this flow is complete terrestrial-non-terrestrial integration in 6G. 6G aims from the start to design terrestrial networks, satellites, and HAPS as a single 3D integrated network, so that a device receives seamless service without being aware of whether it is attached to a ground cell or a satellite. To this end, orbital information sharing, AI-based mobility prediction, and multi-orbit resource coordination are being researched, and NTN is establishing itself as the physical foundation that realizes 6G's vision of universal coverage. However, regulatory and governance challenges—international coordination of satellite frequency and orbital resources (ITU), national sovereignty and regulation, and radio interference coordination—remain variables no less significant than technology.
7. Considerations and Implications
NTN is not merely "adding satellites" but a paradigm that changes the coverage philosophy of communication networks, so from a professional engineer's perspective its application strategy and trade-offs must be examined in multiple layers.
Complement-positioning strategy: NTN should be designed as a complement that fills gaps rather than replacing terrestrial 5G. Given its delay and capacity disadvantages, it is realistic to prioritize application to areas where "connectivity itself is the value"—disaster communications, remote regions, maritime, IoT—and leave urban broadband to the terrestrial network, in a role-division architecture. To maximize effect per investment, apply it selectively starting from areas where the social cost of coverage gaps is large, rather than aiming for uniform nationwide service.
Phased evolution strategy from transparent to regenerative: It is desirable to start early commercialization with a simple transparent payload but follow a roadmap of gradual migration to regenerative payloads and ISL for coverage completeness and capacity expansion. Because a satellite operates for years once launched, a software-defined (SDR)-based flexible design that considers both the standard at launch time and future evolution is key to managing the trade-off.
Realistic management of device and link constraints: The physical limit of an ordinary device's small power and antenna will for the time being constrain the service level to emergency messages and low-rate data. Raising this to broadband requires satellite-side large phased arrays, coverage enhancement techniques, and more satellite constellations, so it is important to manage user expectations with a phased service roadmap without exaggerating the expected level.
Security, sovereignty, and regulatory risk: Because satellites spread signals over a wide area, the eavesdropping, jamming, and spoofing threat surface is broad, and orbital resources and frequencies are subject to international coordination (ITU), which may conflict with national sovereignty and regulation. When applying it to public domains such as disaster communications, review from the perspective of dependence risk and digital sovereignty regarding reliance on specific foreign operators' satellite networks must precede, in addition to encryption and authentication.
Sustainability and the orbital environment: LEO constellations on the scale of thousands of satellites exacerbate orbital congestion, space debris, and light pollution. A sustainable operation strategy that considers safe deorbit after end of satellite life, collision avoidance, and even the carbon cost of launch and operation is required as a long-term challenge.
References
- 3GPP, NTN in Release 17 — https://www.3gpp.org/technologies/ntn-rel17
- Samsung Research, NTN and TN networks for the 6G era — https://research.samsung.com/blog/NTN-and-TN-networks-for-the-6G-era-technology-overview-and-regulatory-challenges
- Cisco, Nonterrestrial Networks: Direct-to-Device Satellite Market White Paper — https://www.cisco.com/c/en/us/solutions/collateral/service-provider/networking/nonterrestrial-networks-satellite-market-wp.html
In one line: NTN integrates non-terrestrial payloads such as LEO·MEO·GEO satellites into the 3GPP standard and ordinary-device ecosystem to fill the terrestrial network's coverage gaps; on top of its service/feeder links and transparent/regenerative payload structures, it solves the unique challenges of large propagation delay, Doppler, and frequent handovers by having the satellite absorb the complexity, realizing 6G's 3D integrated network and universal connectivity while evolving with the trade-offs of inferior delay and capacity and of sovereignty and orbital sustainability.