Rel-18 & Beyond — NTN Evolution

Introduction

NTN evolution is the story of how non-terrestrial networks went from a proof of concept to a production feature of 3GPP. It spans three phases: the Rel-17 baseline that proved NR and IoT could run over a satellite at all; the Rel-18 enhancements that made that baseline seamless and robust enough to deploy; and the Rel-19 / 6G direction that pushes toward capacity, on-board processing and mainstream direct-to-handset service.

This sits at the architectural level rather than the message level: it is about which capabilities land in which release, not the bit-by-bit signalling of any one procedure (those live in the per-topic pages this one links to). Understanding the arc matters because almost every design choice in NTN — transparent vs regenerative payload, how mobility works, which bands and devices are in scope — is really a statement about which release you are targeting.

A word on citation: because the frontier is still moving, this page anchors each capability to a 3GPP release rather than a clause number. Rel-17 and Rel-18 are stable; Rel-19 items are still being shaped and 6G NTN is a research and standardisation direction. Read the later items as trajectory, not as spec you can quote.

Why NTN keeps evolving

Concretely, three pressures drive the evolution. First, link budget: getting an ordinary handheld to close a link to a satellite is hard, so each release squeezes more coverage and uplink capacity out of the design. Second, continuity: a satellite that serves you is useless if your session drops the moment you cross between terrestrial and satellite coverage, so seamless mobility becomes the headline feature. Third, capacity and cost: a bent-pipe satellite that just amplifies wastes a round trip and caps throughput, so the industry pushes toward on-board processing and mainstream, unmodified-handset service. The releases are simply these pressures being answered one at a time.

The Rel-17 baseline — "can it work at all"

It is worth being precise about what Rel-17 actually delivered, because the evolution is defined against it. Rel-17 standardised NTN with a transparent (bent-pipe) payload: the satellite amplifies and forwards, and the real gNB stays on the ground (see ntn-architecture). It supported handheld direct access in FR1 — an ordinary-ish phone talking to a satellite — and VSAT terminals in FR2 for higher-capacity fixed installations. In parallel it delivered the IoT-NTN track (NB-IoT and eMTC over satellite, see ntn-iot). The heavy lifting was in the PHY/MAC: GNSS-based timing and Doppler pre-compensation, ephemeris broadcast, scheduling offsets and scaled timers.

Two constraints shaped that baseline and explain what came next. Because the payload is transparent, the whole radio-protocol stack terminates on the ground, so every radio-loop interaction — HARQ, RRC, scheduling — pays the full feeder-plus-service-link round trip; the satellite is a mirror, not a node. And because the earliest platforms emphasised GEO and delay-tolerant operation, the baseline optimised for reachability and coverage over throughput and latency. Both are exactly the constraints the later releases attack.

Rel-18 — enhancing the baseline

Rel-18 keeps the transparent-payload architecture but sharpens almost every rough edge Rel-17 left. The headline enhancements:

• Coverage and uplink capacity / link budget improvements for handheld devices, so an ordinary phone closes the link more reliably.
• Improved network-verified UE location, tightening the regulatory-grade positioning the AMF depends on (see ntn-5gc).
• NTN–TN and NTN–NTN service continuity and mobility — moving between terrestrial and satellite coverage, and between satellites/constellations, without dropping the session.
• Support for bands above 10 GHz — deployment and coexistence considerations for higher-frequency NTN operation.
• Disaster roaming / resilience, letting devices fall back to satellite access when terrestrial networks fail.
• Enhancements to discontinuous coverage and power saving, so a UE in a sparse constellation predicts coverage gaps and sleeps through them.
• Store-and-forward direction and early steps toward regenerative payloads — putting an on-board gNB (or part of one) on the satellite rather than on the ground.

Rel-18 also carried the IoT-NTN track forward in parallel with NR-NTN — improving mobility and service continuity for NB-IoT/eMTC devices, handling discontinuous coverage for sparse constellations, and easing the GNSS burden so a battery-bound device re-acquires its position (and therefore its timing/frequency pre-compensation) less often. The two tracks evolve together but keep their distinct radios and cores.

NTN–TN mobility — why it is the big one

Of all the Rel-18 items, terrestrial↔non-terrestrial mobility is the one that changes the user experience most, so it deserves its own section. In the Rel-17 world a UE was either on the ground network or on the satellite; moving between them was not a smooth, session-preserving affair. NTN–TN service continuity makes satellite coverage an extension of the terrestrial network rather than a separate island: a phone that walks out of terrestrial coverage can hand over to a satellite, and back again, keeping its PDU session alive.

NTN–NTN mobility is the companion piece — handing a UE between satellites, and even between different constellations, as they pass. Together these turn a patchwork of coverage into one continuous network (the mechanics live in ntn-mobility). This is what "ubiquitous coverage" actually means in practice: not just that a satellite can serve you, but that you cross the seam without noticing.

Rel-19 and beyond — the forward look

Past Rel-18 the picture is genuinely evolving, so read this as direction rather than done deals. Several threads are gathering momentum:

• Regenerative payloads maturing — moving from a bent pipe toward an on-board gNB that processes on the satellite, cutting the feeder-link round trip out of the radio loop and enabling inter-satellite routing.
• Direct-to-device / "direct-to-cell" momentum — unmodified smartphones talking to satellites, increasingly using MSS (mobile-satellite service) spectrum, is the commercial centre of gravity.
• Broadband NTN and higher throughput, pushing beyond messaging and voice toward real data rates.
• RedCap over NTN, bringing reduced-capability mid-tier devices (wearables, industrial sensors) to satellite.
• Groundwork toward 6G NTN and fully integrated terrestrial–non-terrestrial networks, where satellite is a native layer of the system rather than an add-on.

The regenerative payload is the pivot the rest hinges on. Once an on-board gNB terminates the radio protocol on the satellite, the radio loop no longer pays the feeder-link round trip — HARQ, scheduling and RRC see only the service link — and satellites can route to each other over inter-satellite links rather than dropping every packet to a ground gateway first. That is what makes broadband NTN and low-latency direct-to-cell plausible rather than merely possible. It also reframes RedCap-over-NTN as a natural next tier between full NR handsets and the frugal IoT-NTN devices.

Be honest about the status: Rel-19 items are still being shaped, and 6G NTN is a research and standardisation direction, not a specification you can quote. What is clear is the destination — on-board processing, mainstream direct-to-handset, and a network that treats ground and space as one fabric.

Rel-17 vs Rel-18 vs Rel-19+ at a glance

The four dimensions that capture the trajectory — payload, mobility, bands, and target device — line up cleanly per release.

Summary

The NTN arc is three clean phases. Rel-17 proved the concept: a transparent (bent-pipe) payload with the real gNB on the ground, handheld direct access in FR1, VSAT in FR2, and a parallel IoT-NTN track, all built on GNSS pre-compensation, ephemeris broadcast, and scaled timers/offsets. Rel-18 made that baseline deployable — better handheld link budget and uplink capacity, network-verified location, bands above 10 GHz, disaster-roaming resilience, discontinuous-coverage power saving, and, above all, seamless NTN↔TN and NTN↔NTN mobility that keeps a PDU session alive across the seam. Rel-19 and 6G point at capacity and cost: regenerative on-board processing, inter-satellite routing, broadband and RedCap over NTN, and mainstream direct-to-device on MSS spectrum.

The single throughline: NTN moved from "can it work at all" to "make it seamless and robust" to "make it a native layer of the network." Cite each capability by the release that carries it, hold the Rel-19/6G items as direction rather than settled spec, and the whole landscape stays legible.