10 What the phone says to the network

10.1 Why there is a whole layer after the architecture

The architecture, phase by phase — and backhaul, which runs the other way ends with a decision: this box knows that, that box asks this one, this rule applies here. None of that is yet something a device can send.

Turning it into messages is the CT groups' job. They write the protocol a device speaks to the network — every field, every value, every error, and what happens when a message does not arrive.

It is a large block of work. CT1 carries 2164 documents of this subject over 37 of its meetings, running from meeting C1-126-e to meeting C1-162, and outside the radio groups only SA2 carries more.

10.2 Phase 1: which network should a phone even choose?

5GSAT_ARCH-CT is the Release 17 item, led by CT1, sitting on five work plan rows, planned from 2020-09-16 to 2022-03-11, and carrying 912 documents over 92 meetings ^[4].

Its groups are the CT plenary, CT1, CT3, CT4 and CT6 — five of the six CT bodies in this register at once 5GSAT_ARCH-CT.

Its justification is the plainest statement in this whole register of how 3GPP work moves down the chain:

Requirements, then architecture, then this. Two sentences and the whole order of work is in them.

The item then does something unusual: it opens with a study of its own, run by one group.

PLMN selection is how a device decides which network to attach to. On the ground that question is answered by which masts a device can hear and what its subscription allows.

Over a satellite the coverage may cross a border, so the same question needs a new answer — and it needs one before any message can be designed.

The report that came out of it is TR 24.821, which carries 213 documents, exists at one version and had its newest file uploaded on 2021-09-25. Its register entry records no parsed text, so this school can name it and not quote it.

The item's change table then names eight documents, and the two that matter most here are TS 24.501, what a device says to the 5G core, and TS 24.301, what it says to the older one.

The wording of the entries is worth reading: updates related to satellite access for the 5G one, and potential updates for interworking between the two cores for the older one 5GSAT_ARCH-CT. One is a commitment, the other is a maybe.

10.2.1 What came of the network-selection study

The item's change table says where the answer was meant to go. Beside TS 23.122 — the document that says how a device in idle mode picks a network — it writes "updates to PLMN selection related to satellite access depending on outcome of study phase" 5GSAT_ARCH-CT.

That document carries 173 documents of this subject.

So the chain runs: a requirement in TS 22.261, an architecture conclusion in TR 23.737, a CT study in TR 24.821, and a promise of edits to a document that existed long before any of this. None of it is visible from the work item's short name alone.

10.3 The two items that are CT-led but do not look it

Two large items sit in CT's lists and are easy to file under the wrong group.

IoT_SAT_ARCH_EPS is the Release 17 item for architecture support for the narrowband internet-of-things radio and its enhanced machine-type twin in the older packet system.

Its leadership line names SA2, but its groups are CT1, CT3, CT4, CT6 and SA2 together, it sits on five work plan rows, and it carries 518 documents over 75 meetings ^[5] IoT_SAT_ARCH_EPS.

Its documents run from meeting C1-132-e to meeting CP-111 — a CT working group at one end and the CT plenary at the other.

5GSAT_Ph2 is phase 2, the architecture item of The architecture, phase by phase — and backhaul, which runs the other way. Its leadership line names SA2, and its groups are CT1, CT6 and SA2 5GSAT_Ph2.

The newest of its 810 papers in this record sits at meeting C1-160, which is a CT1 meeting ^[6].

10.4 Phase 3: nineteen documents

5GSAT_Ph3_ARCH is the Release 19 item, again led by CT1, sitting on four work plan rows, planned from 2024-09-09 to 2025-09-09, carrying 862 documents over 43 meetings ^[7]. Its groups are CT1, CT3 and CT4 5GSAT_Ph3_ARCH.

Its objective is short and says nothing about satellites at all:

That is the honest shape of a protocol work item. It does not decide anything; it promises to write down whatever the architecture groups decided, and to wait for them where they are not finished.

What makes the item large is the list underneath. Its change table names nineteen documents 5GSAT_Ph3_ARCH, and they fall into three groups:

• P1. What the device sends. TS 24.301 for updates to the signalling procedures for store and forward operation, TS 24.229 for the telephony side of two devices talking through one satellite, and TS 27.007 for the commands a modem accepts.

• P2. What one core network function says to another. Fifteen of the nineteen are numbered in the 29 range, including TS 29.571 for the shared data types every one of those interfaces uses, and three of them for telling an outside application that a stored message is still on its way.

• P3. One document outside both groups: the short message service specification, TS 23.040. It is not in this register, because nothing else in this subject points at it.

Nineteen documents for one release of one feature is what "the protocol work is bigger than the architecture work" means in practice.

10.5 Phase 4, which has barely started

5GSAT_Ph4-CT is the Release 20 item, led by CT1, sitting on four work plan rows, and it has barely begun: one document, one meeting, a completion value of 0.03, created on 2026-06-03 and planned to finish 2027-03-03 5GSAT_Ph4-CT.

Its own describing document is CP-261281, titled "New WID on CT aspects for integration of satellite components in the 5G Phase 4", filed by CT1 at meeting CP-112 and uploaded on 2026-06-10.

Its outcome word in the store is "approved", which is a fact about that document and not about the work CP-261281.

Its change table already names twelve documents, and reading them tells you what phase 4's protocol work will be about: the telephony call control documents, the device's own protocol for the older core, and the modem commands.

After those comes a run of core network interface documents, each entry saying the change is "to support a non-IP mechanism to transport voice packets" 5GSAT_Ph4-CT.

Put beside The architecture, phase by phase — and backhaul, which runs the other way, the picture is consistent.

The architecture study opened on a voice call over a geostationary satellite through the narrowband internet-of-things radio, and this is the protocol side of the same sentence — the same release, with the architecture row created on 2025-12-03 and this one on 2026-06-03.

10.6 The other CT groups, and the plenary

CT1 is not alone. Four more CT bodies carry documents of this subject.

Group	Documents	Meetings	First and last meeting here
CT1	2164	37	C1-126-e to C1-162
CT3	202	28	C3-113-e to C3-146
CT4	180	30	C4-101-e to C4-136
CT6	73	16	C6-106-e to C6-127
CT	144	23	CP-90-e to CP-112

The register lists those groups by name and does not say what each one is for in general, nor which of them still meet. What it does hold is which work items each group's documents carry, and that is all the description below claims.

• CT3's documents here carry the phase 1, phase 2 and phase 3 items, backhaul, the internet-of-things architecture item and the phase 3 application-layer item CT3.

• CT4's carry the same phases plus the Release 19 NR phase 3 radio item CT4. CT4 also leads the Release 20 backhaul quality-of-service protocol item, which carries no document at all, stands at a completion value of 0, and names one document to change TEI20_QoS_MWAB_SAT-CT.

• CT6's carry phase 1, phase 2, the phase 1 feature row and the internet-of-things architecture item, and CT6 is named in the leadership line of the internet-of-things conformance item CT6.

• The CT plenary carries 144 documents over 23 meetings and is where each of these items was opened CT — every CT work item in this chapter names a CP document as the paper that describes it. Work is opened at a plenary and done in the working groups, and its count is small beside CT1's.

10.7 One item that is CT-led and easy to file wrongly

PWS_NTN — public warning messages over satellite radio — is not an SA item despite what its subject suggests. Its leadership line names CT1, its groups are CT1 and CT4, it sits on three work plan rows, it carries 60 documents over six meetings, and its change table names TS 23.041 PWS_NTN.

It is described in What the system was asked to do beside the other small single-purpose items, because that is where a reader looks for it. It belongs here by ownership.

10.8 Where the two protocol documents stand

Everything above ends up in two places more often than anywhere else.

• TS 24.301 — what a device says to the older packet core. 705 documents of this subject, 704 of them change requests, and three work items name it in their own change table: all three CT phases above.

• TS 24.501 — the same for the 5G core. 560 documents, 559 of them change requests, and exactly one work item names it in a change table, the Release 17 one.

That asymmetry is worth a second look. The newer core's protocol document is named by one change table; the older core's is named by all three.

The wording of the phase 4 entry says what it is there for — enhancements to support a voice call and an emergency call over the narrowband internet-of-things radio through a geostationary satellite 5GSAT_Ph4-CT.

Nearly every document against either specification is a change request rather than a discussion paper. The discussion sits in the study reports and in the architecture work; what arrives here is the edit.

10.9 Where this sits in the rest of the course

• R1. The architecture, phase by phase — and backhaul, which runs the other way — the decisions this work writes down.

• R2. TS 24.501 and TS 24.301 are the two documents to open if you have to know what a device actually sends. This school carries the text of neither.

• R3. CP-261281 is the shortest statement of what CT promised for phase 4, and it is one page rather than a specification.

• R4. The overview course gives the CT side a section, and the quick-start course does not go into it at all.