Cheat sheet

IMS — the call as an internet session

The working reference on 3GPP's IMS work, from Release 5 to Release 20

Depth: In depth
Chapters: 14
Reading: 165 min
Questions: 69
Sources cited: 399

Back to the course

The shape of it

1 What IMS is, and how this subject is bounded

2 How it grew, release by release

3 Who does the work

4 What is written down

5 Where it stands, and reading the record yourself

Keep these in your head

1.1 What IMS is, in the words of its own documents
There is no single IMS document. Four specifications, owned by three groups, each answer a different question about the same system — requirements, architecture, protocol, call model. Deciding which of the four you are standing in is the first move in reading anything about IMS.
2.1 How this subject is cut out of the work plan
Every number in this course is a count over rows chosen by one rule: the acronym of the row carries the letters IMS with no capital letter in front of them. Change the rule and every number changes. Nothing was chosen by reading a title, and nothing by hand.
3.1 Release 5: the founding, and the ring of work around it
The founding row of IMS is one row, acronym IMS, title "Provisioning of IP-based multimedia services", marked Rel-5 and listing eight groups. It is the only row in the whole subject whose acronym is the bare three letters.
4.1 Releases 6 and 7: interworking, messaging, and a stream that never stopped
Rel-6 is where IMS was made useful rather than complete: interworking with the circuit-switched world so a session could reach an ordinary telephone, plus messaging and group management. Rel-7 opened the IETF protocol alignment work, which was still being reopened under a new number for Rel-20.
5.7 Releases 8 and 9: emergency calls, continuity and the fixed-network merge
"Common IMS" is the work plan's name for the specifications brought into 3GPP from TISPAN and maintained there, and three rows exist only to document and maintain them. A row whose whole job is looking after somebody else's text is ordinary in a mature area and easy to mistake for new work.
6.3.1 Releases 10 to 14: the browser, telepresence and the middle years
A browser client did not get a private corner of IMS. It got a stage 3 specification of its own, TS 24.371, and then took its place among the ordinary documents that the ordinary maintenance work keeps changing.
7.3 Releases 15 to 17: IMS moves onto the 5G core
Rel-15 did not redesign IMS for 5G. It changed the documents that had to be changed for an IMS call to work when the pipe underneath it was a 5G core — and said in writing that the 5G core's own problems were somebody else's work item.
8.4 The map: fourteen groups, 309 meetings and 374 names
The shape of a document count map tells you where a subject really lives. IMS puts its three largest counts in the CT working groups, which write the bytes on the wire. That is a core-network subject and not a radio one, and the map says so before you read a word of any document.
9.2.1 CT1, CT3, CT4 and CT6: where most of it happens
One document explains CT1's whole row on the map. TS 24.229 is the protocol every IMS implementation has to follow, and 1680 documents in this record exist to change it. A specification everybody implements attracts change requests the way nothing else in 3GPP does.
10.2 SA1 to SA5, the two RAN groups, and the plenaries above them
Stage 2 is one document and one group. If you want to know what IMS is rather than what it says on the wire, TS 23.228 is the document and SA2 is the group that holds the pen on it.
11.2 The specification family: 241 documents, and the two that are IMS
The list of 241 specifications is built from what work items declared and from where change requests actually landed. It is not a list of documents with IMS in the title, and it is not a list of documents any group owns.
12.2 Maintenance is the main event: change requests and protocol alignment
A change request is not a bug report and not a release note. It is one proposed edit to a published document, written as the exact old text and the exact new text, brought to one meeting and dealt with there. Every rule in IMS that is not in its first version arrived this way.
13.3.1 What is running now: Releases 18 to 20
Resiliency is the Release 19 story, and it was not proposed as an improvement. It was proposed because real networks fell over, and the proposers named the industries that were hit. Its justification reads as an incident report rather than as a proposal.
14.2 Reading the record yourself, and what it cannot tell you
Search by acronym, never by title. The acronym is what documents are tagged with; the title is free text that often does not contain the word you are looking for. This one habit is the difference between finding a subject and finding a handful of it.

Easy to get wrong

One from each chapter that has one. The course carries 17 in all.

1.2 What IMS is, in the words of its own documents
"IMS" and "voice over IMS" are not the same subject. TS 22.228 is about multimedia sessions in general; telephony and its supplementary services — call waiting, diversion, the things a telephone exchange used to do — live in TS 22.173. A sentence about IMS that is really about a phone call is a sentence about one service that IMS carries.
2.3 How this subject is cut out of the work plan
Never read the row count as a count of things people worked on, and never read the acronym count as one either. 345 is rows, 177 is acronyms, and the number of pieces of work a person would name out loud is neither of them — no field in the plan holds it.
3.3 Release 5: the founding, and the ring of work around it
The description S1-000299 is quoted here on a chain of reasoning, not on a filing record. The store has no row for that document number; the file was taken as this work item's own because its own title, "Support of IP multimedia services", shares three words with the work plan row's name. The quote is honest, the attachment is inference, and the register says so in its own method line rather than hiding it.
4.2 Releases 6 and 7: interworking, messaging, and a stream that never stopped
Interworking with circuit networks and interworking with IP networks are two different work items with two different documents, and people mix them up constantly because both start IMS-CCR-IW. IMS-CCR-IWCS and TS 29.163 are the circuit side; IMS-CCR-IWIP and TS 29.162 are the IP side. Checking which of the two an argument is about costs one glance and saves an hour.
5.1 Releases 8 and 9: emergency calls, continuity and the fixed-network merge
It is easy to write that Rel-8 and Rel-9 are the two biggest releases of IMS work, and it is wrong. Rel-8 leads with 64 rows, but Rel-12 has 60 and Rel-9 has 45, so Rel-12 sits between them. Any "biggest" here is also bounded by the acronym rule of this register and says nothing about 3GPP work that rule does not reach.
6.2 Releases 10 to 14: the browser, telepresence and the middle years
Those two columns answer different questions and must never be put in one sentence as if they answered the same one. Rel-12 carries 60 work plan rows and 433 documents; Rel-13 carries 15 rows and 830 documents. Rows count entries a plan keeps, documents count paper filed against a release, and the document column is a floor rather than a total — this store holds no IMS document that carries a work item acronym before 2015.
7.5 Releases 15 to 17: IMS moves onto the 5G core
543 is a count of documents whose work item column names eIMS5G_SBA, in this register, on the day it was read. It is not a measure of how large the change was, how long it took or how hard it was argued. A count filtered by acronym also misses every document that carries no acronym at all, so it is a floor.
8.4.3 The map: fourteen groups, 309 meetings and 374 names
A group with no entry here is not a group that was not involved, and a group with one document is not a group that did one thing. This map counts documents that carry one of the 177 IMS work item acronyms ^[1]. Everything filed before an acronym existed to file it under, and everything filed without one, is missing from every row.
9.2.4 CT1, CT3, CT4 and CT6: where most of it happens
The last meeting in CT1's span has not happened. C1-162 at Prague runs from 2026-08-24 to 2026-08-28 and its state is recorded as not started, with three IMS documents already filed against it C1-162. A "first and last meeting" pair describes the span of the stored rows, not a stretch of work that has finished.
10.6 SA1 to SA5, the two RAN groups, and the plenaries above them
That single RAN2 row may not be called a change request, a liaison or work of RAN2's own. The register records the acronym and the meeting and does not record the document's kind. One row is exactly as much as the record supports, and naming its kind would be an invention that reads as a fact.
11.8 The specification family: 241 documents, and the two that are IMS
A document with a 23-, 24- or 29-number is not "an IMS document" or "a 5G document" by its number. TS 29.562 belongs to the 5G core and carries 230 IMS papers; TS 23.218 is an IMS document that no change request in this record touches. Read the counts, never the numbering.
12.3 Maintenance is the main event: change requests and protocol alignment
"1736 documents were agreed" is not "1736 things were decided". A document revised three times is three rows, each with its own word, all about the same edit. Never add the outcome words up and never read one as a verdict on the work behind it.
13.8 What is running now: Releases 18 to 20
Nothing on this page may be called finished, running or stopped, and this course does not call it any of those. The completion percentage is typed in by hand — 339 of the 345 rows read 100% ^[2] — the word Deleted is typed into a title, and a document's outcome word is about that document alone. Where somebody wants a status, the answer is the newest documents under the acronym and the current version of the specification, not a field.
14.5 Reading the record yourself, and what it cannot tell you
Two numbers in the same sentence are usually two different questions. "1680 documents change TS 24.229" and "241 specifications are touched" are both true and cannot be combined into anything ^[3] ^[4]. Whenever a comparison feels natural, open both method lines first — that is exactly the moment this data misleads.

The numbers, and where each came from

Every one was counted out of a file on this machine. Hover a row to read the question that produced it.

Where to look it up

The 281 named sources it cites, by what they are.

Work item 93

eIMS5G_SBA FS_enIMS_ARC IMS IMS-CODEC IMS_SSFDD IMS_Corp2 IMS-ASEC IMS-LI IMS-CCR IMS-CCR-IWMM IMS-CCR_Test IMS-CAMEL IMS-OAM IMS-ONOSA IMS-TEST IMS-FrWk IMS-Sex IMS_CONV_IP_COM-UEConTest IMS-CCR-IWCS IMS-CCR-IWIP IMSM IMSGM IMS2 IMSProtoc IMSconf IMSProtoc2 IMSProtoc20-CT MIW-IMS IMS_Comm_Mtce7 IMS_EMER_GPRS_EPS IMS_SCC eIMS_RP IMS-Sec IMS_IBCF IMS_Corp IMS_Comm_Doc IMS_Comm_Mtce8 IMS_EMER_GPRS_EPSC1 IMS_EMER_GPRS_EPS-Sec IMS_EMER_GPRS_EPS-SRVCC IMS-Cont IMS_WebRTC IMS_WebRTC-CT IMS_WebRTC-CT1 IMS_WebRTC-CT4 IMS_TELEP IMS_TELEP_S4 IMS_TELEP_EXT NWK-PL2IMS NWK-PL2IMS-CT IMS_RegCon IMS_RegCon-CT IMS_RegCon-del1 IMS_SDE SDPCN_IMS AT_IMS EHNB-IMS IMS-UICC-S FS_IMS_P2P FS_eIMS IMS_SC_eIDT CC_IMS_USIM IMS_UApAs IMSSat IMSProtoc4 IMSProtoc5 IMSProtoc6 IMSProtoc7 IMSProtoc8 5GS_Ph1-IMSo5G FS_eIMS5G FS_eIMS5G2 FS_enIMS enIMS SCAS_IMS IMS_RES-CT SR_IMS AIML_IMS-MED FS_IMSRE_SEC IMSDCDataOff CH8_IMS 5GS_Ph1_IMSCH IMSProtoc17_dataCH-UEContest IMSProtoc3 IMSProtoc9 IMSProtoc16 IMSProtoc17 IMSProtoc18 IMSProtoc19 FS_IMS_RES FS_IMSResil VoIMS IMS-MMTel

Quoted wording 69

the scope of 23.228 the scope of 24.229 "IM CN subsystem" as 22.228 defines it "Conference" as 22.228 defines it "Data channel application" as 23.228 defines it the scope of 22.228 the scope of 23.218 the scope of 22.173 "IP multimedia session" as 22.228 defines it "IP multimedia application" as 22.228 defines it "IP multimedia service" as 22.228 defines it "Access independence" as 22.228 defines it "IMS application" as 23.228 defines it "Application data channel" as 23.228 defines it what IMS set out to do why IMS was proposed what IMS-CODEC set out to do why IMS-CODEC was proposed what IMS-CCR-IWCS set out to do why IMS-CCR-IWCS was proposed what IMSM set out to do why IMSM was proposed what IMSGM set out to do why IMSGM was proposed what IMS_EMER_GPRS_EPS set out to do why IMS_EMER_GPRS_EPS was proposed why IMS_SCC was proposed why eIMS_RP was proposed what IMSProtoc2 set out to do why IMSProtoc2 was proposed what IMS_Corp set out to do the scope of 23.167 the scope of 23.292 the scope of 23.237 what IMS_SCC set out to do what eIMS_RP set out to do the scope of 23.380 what IMS-Sec set out to do what IMS_IBCF set out to do what IMS_WebRTC set out to do what IMS_TELEP set out to do why FS_eIMS was proposed why IMS_WebRTC was proposed why IMS_TELEP was proposed why 5GS_Ph1-IMSo5G was proposed what 5GS_Ph1-IMSo5G set out to do why FS_eIMS5G was proposed what eIMS5G_SBA set out to do why FS_eIMS5G2 was proposed what SCAS_IMS set out to do what FS_eIMS5G set out to do why eIMS5G_SBA was proposed what FS_eIMS5G2 set out to do what FS_enIMS set out to do what enIMS set out to do why SCAS_IMS was proposed the scope of 23.334 why SR_IMS was proposed why IMSProtoc20-CT was proposed why FS_IMS_RES was proposed what FS_IMS_RES set out to do what IMS_RES-CT set out to do what SR_IMS set out to do why FS_IMSResil was proposed what FS_enIMS_ARC set out to do why FS_enIMS_ARC was proposed why FS_IMSRE_SEC was proposed why AIML_IMS-MED was proposed the scope of 23.801-04

Specification 69

TS 22.228 TS 23.228 TS 24.229 TS 23.218 TS 22.173 TR 23.820 TS 29.163 TS 24.228 TS 29.061 TS 29.162 TR 23.821 TR 24.930 TS 24.147 TS 23.167 TS 23.401 TS 23.402 TS 22.101 TR 23.869 TS 23.292 TS 23.237 TS 23.380 TS 29.238 TS 29.235 TS 23.204 TS 23.506 TS 23.521 TS 23.406 TS 23.417 TS 24.371 TS 24.103 TS 26.223 TS 29.165 TS 26.114 TS 23.003 TS 23.008 TS 24.167 TS 29.213 TS 29.214 TS 29.228 TS 29.229 TS 29.328 TS 29.329 TR 23.794 TS 23.501 TS 23.502 TS 23.503 TS 29.510 TS 29.513 TS 29.514 TS 29.571 TR 23.700-10 TR 23.700-11 TR 23.700-12 TR 22.823 TS 33.226 TR 33.926 TS 29.562 TR 29.866 TR 29.867 TS 23.334 TS 26.567 TS 34.229-1 TS 26.264 TS 24.186 TS 29.292 TR 23.801-04 TR 29.949 TR 33.768 TR 23.801-03

Meeting document 23

SP-000216 SP-000398 RP-171635 SP-020065 CP-060101 SP-080844 SP-080633 CP-091050 CP-080794 C4-244485 C1-246921 C3-250657 S2-2413007 SP-251237 CP-260044 CP-232281 CP-253168 SP-240492 CP-261069 SP-260686 SP-250878 C4-261402 C1-262608

Working group 14

SA2 CT1 CT4 CT3 CT SA4 RAN5 SA3 SA1 SA5 SA RAN CT6 RAN2

Meeting 13

CP-112 C1-162 C1-112 C1-111 C1-90 C1-91 C1-92 C1-95 C4-68 CP-110 CP-111 SP-112 SP-108

Where to go from here

The same subject, quick start — Under an hour, four chapters, the shape of the thing.
The same subject, overview — Enough to follow the conversation and know what to look up.
Every question of this course — 69 of them, on one page.
The final test — 15 questions, 70 per cent to pass.
Every source behind the course — 858 of them, each with its method.

The raised numbers on this page

177 different work item acronyms carry IMS — distinct values of the acronym field over those 345 rows, read 2026-08-05
339 of the 345 IMS work plan rows stand at 100% — rows taken in by the acronym rule whose completion field reads 1, read 2026-08-05. The field is a percentage the work plan keeps by hand and it is not a status word
1680 IMS documents name 24.229 as the document they change — IMS documents whose change-request column names 24.229, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05, of which 1675 are of type CR, CR pack or draftCR
241 specifications this work touches — catalogue rows for every document a work item names as one it changes, plus every one named in the change-request column of at least 3 IMS documents, asked of /var/www/whatthespec.net/data/database/api/spec_catalog.sqlite on 2026-08-05