7 SA2 and the specification that did not exist

7.1 What an architecture group is for

Stage 1 says a battery-less device must be reachable. Somebody then has to decide what exists in the network to reach it: which functions, which interfaces, which identifiers, which procedures run in which order.

That is stage 2, and in 3GPP it belongs to SA2. It draws the boxes and the arrows. It does not define a single byte on the wire — that comes later, from the CT groups.

For Ambient IoT this meant something unusual. The existing 5G architecture document could not simply be amended, because what was being described did not fit inside it. A new specification had to be written.

7.2 The study first

SA2's study is FS_AmbientIoT, unique id 1020071 — the same short name SA1's study carries, a different item entirely FS_AmbientIoT FS_AmbientIoT. It was approved as SP-240969 at plenary SP-104 SP-240969, and it says plainly where it takes its input from:

That single sentence is the whole handover from SA1 and the requirements nobody designed against yet. The study ran to 2024-08-30, stands complete, and its report is TR 23.700-13, against which 746 documents were filed FS_AmbientIoT TR 23.700-13.

7.3 Then the specification

The normative work is AmbientIoT-ARC, unique id 1070010, approved as SP-250377 at plenary SP-107 AmbientIoT-ARC SP-250377. Its objective is precise about scope in a way that is easy to skim past:

Two things in that sentence decide how much of the problem was actually solved in Rel-19:

• B1. It builds only on the conclusions of the study, clause 8 — not on every solution the study explored. A study report typically holds several candidate answers; one is chosen.

• B2. It focuses on Device 1 in Topology 1. Those are the terms the documents use for one device type and one connectivity arrangement. What they mean is settled in TS 38.300, which defines the device types, traffic types, use cases and topologies ^[3] TS 38.300 — and this school does not carry that text.

The work stands complete as of 2025-09-09 and carries 695 documents AmbientIoT-ARC.

7.4 What TS 23.369 is

TS 23.369, architecture support for Ambient power-enabled Internet of Things, stage 2, is the document that did not exist before this work TS 23.369. Its scope states both what it covers and the boundary that surprises people:

The specification also had to invent vocabulary, because there was none. Two of its definitions carry the shape of the whole design:

An area with its own identifier, holding several radio nodes and several readers, is not how a 5G cell works. It is a shape built for a device that is not attached to anything and may be reachable from more than one direction at once.

7.5 What SA2 changed in the documents that already existed

Two existing SA2 specifications were amended so the new one could plug in:

• TS 23.501 TS 23.501 — the architecture of the 5G system itself. 14 Ambient IoT change requests.

• TS 23.502 TS 23.502 — the procedures of the 5G system. 26 Ambient IoT change requests.

Fourteen changes to TS 23.501 is a small number for a feature of this size, and it fits what the scope said: in this release the Ambient IoT system is an isolated private network of its own, so most of it does not have to be wired into the public 5G architecture at all ^[3].

7.6 Why a new specification rather than an amendment

3GPP does not create stage 2 specifications lightly. TS 23.501 and TS 23.502 between them describe the whole 5G system, and most new features arrive as change requests to those two.

Ambient IoT did not. It got a document of its own, and the reason is visible in the numbers: TS 23.369 took 524 change requests while TS 23.501 took 14 and TS 23.502 took 26 TS 23.369 TS 23.501 TS 23.502.

Two small amendments to the existing architecture, and a document of its own that took more change requests than anything else in the area. That is not a feature being added to the 5G system.

It is a second system described beside it, with a few hooks where the two meet — exactly what the "isolated private network" sentence says ^[3].

7.7 Phase two: widening the shape

Phase 2 began before phase 1 finished. The study FS_AmbientIoT_Ph2_ARC, unique id 1080063, opened as SP-250834 at plenary SP-108 FS_AmbientIoT_Ph2_ARC SP-250834, and its first work task says what it is after:

The study stands complete, finished 2026-03-03, and its report is TR 23.700-30 with 264 documents against it FS_AmbientIoT_Ph2_ARC TR 23.700-30.

The normative follow-on is AmbientIoT_Ph2-ARC, unique id 1110030, opened as SP-260296 at plenary SP-111, and it is running, not finished AmbientIoT_Ph2-ARC SP-260296. What its recorded progress does and does not mean is Where the work stands, and what a percentage is worth.

7.8 The letter-writing

SA2 sent 139 letters to other groups SA2. The area sent 351 in total ^[7], and the next two senders — SA3 and RAN2 — managed 38 each SA3 RAN2.

Two recent examples show the kind of thing being settled by letter in mid-2026: a reply about sharing a correlation identifier S2-2605244, and a response about the phase 2 report's conclusions S2-2605226. Both are from SA2 at meeting S2-175.

What either letter actually said is not something this school holds. The register gives the number, the meeting, the sender and the address on 3gpp.org, and the reading is yours.

7.9 Where to read on

Three documents, none of whose text this school carries:

• TR 23.700-13 TR 23.700-13 — the phase 1 architecture study. Clause 8 is the one the work item points at.

• TS 23.369 TS 23.369 — the specification itself. Start here if you want the system rather than the history.

• TR 23.700-30 TR 23.700-30 — the phase 2 study, for where it is going next.

The protocols that carry all of this are CT1, CT3, CT4 and CT6 — one work item description, four groups; the security model around it is SA3, and why a device with no battery is a hard security problem.