School of Specs IMS — the call as an internet sessionIn depth

Where it stands, and reading the record yourself · chapter 13 of 14 · 12 minutes

13 What is running now: Releases 18 to 20

The present tense of this subject — resiliency after real outages, split rendering, an architecture study aimed at 6G-era services, and why none of it may be called finished.

13.1 The hardest question this record answers

"What is happening now" sounds like the easiest question to ask of a document store and it is the hardest, because the store holds no status of any kind. There is no field saying a piece of work is running, stopped or done. There is no date for any release. There is not even a reliable finishing day.

What there is: work plan rows marked with a release, documents marked with a release, papers that opened each piece of work, and the words the proposers wrote. This chapter reads the present tense out of those four and says plainly where each of them stops.

The newest IMS document in this record went up on 2026-08-04 [9], and nothing on this page reaches past that day.

13.2 The size of the near end of the record

Three release markers cover what is recent.

Release Work plan rows Documents
Rel-18 4 [10] 217 [11]
Rel-19 10 [12] 387 [13]
Rel-20 7 [14] 189 [15]

A row is one group's share of one work item, so a handful of acronyms fills those numbers: a work item that CT1, CT3 and CT4 all work on is three rows and one job.

A document's release marker is the release it was filed against, and nothing more. It does not date the document and does not say the release is open.

Release 18 is thin here: 4 rows [10], and the protocol alignment turn described in Maintenance is the main event: change requests and protocol alignment accounts for 3 of them, across CT1 and CT3, with 45 documents IMSProtoc18 [16].

One document first appears at Release 18 — TS 24.186, the data channel applications protocol, whose earliest stored version is Rel-18 at 18.5.0 TS 24.186.

The filing years are the other view of the same end of the record. 209 IMS documents were filed in 2024 [17], 365 in 2025 [18] and 166 so far in 2026 [19].

Those are counts of when files went up, they leave out the 237 documents with no upload time at all [20], and the last of them covers part of a year. Read them as a shape, never as a rate.

13.3 Release 19: IMS falling over

The centre of Rel-19 is not a new service. It is the discovery that the thing breaks, and the reason is written down without any softening.

That is the justification of FS_IMS_RES, the CT4 study on disaster prevention and restoration, opened by CP-232281 FS_IMS_RES CP-232281. It drew 147 documents [21] and its report is TR 29.866 TR 29.866.

Its own objective is as plain as its justification.

13.3.1 What followed the study

A study is homework and binds nobody, so the work that matters is the work item that came after it. IMS_RES-CT does exactly one thing.

It sits across CT1, CT3 and CT4, carries 101 documents [22], was opened by CP-253168 CP-253168, and declares sixteen documents it will change IMS_RES-CT.

Two of them are the restoration procedures and the IMS-ALG to access-gateway interface TS 23.380 TS 23.334, and several more are 5G core documents rather than IMS ones.

That spread is the point. Making IMS survive a failure means changing the 5G core's policy and subscriber-data interfaces too, which is why the work needs three groups.

13.3.2 The other Release 19 item

SR_IMS put split rendering into the IMS architecture — the idea that some of the work of drawing a picture happens in the network rather than on the device. SA4 owns it, it carries 84 documents [23], and it was opened by SP-240492 SR_IMS SP-240492.

Its reason is a piece of 3GPP housekeeping rather than a new idea. Release 18 had already specified split rendering over the generic real-time architecture, and workflows for split rendering over IMS had been specified inside another work item.

SR_IMS gathered that into the IMS architecture proper [24]. Its own document is TS 26.567 TS 26.567.

13.4 Release 20: seven rows

Rel-20 has 7 work plan rows [14] and five acronyms worth knowing.

FS_IMSResil carries the resiliency question forward, and its justification describes a very specific vicious circle.

A device that gets no answer asks again; enough devices asking again is what turns a fault into an outage.

The study is CT4's, carries 93 documents [25] and writes TR 29.867 TR 29.867. The paper the work plan points at is CP-261069, a revised study item approved at the CT plenary's 112th meeting CP-261069, and its completion field reads 70% FS_IMSResil.

FS_enIMS_ARC is the architecture study, SA2's, with 51 documents [26], pointed at by SP-260686 SP-260686 and standing at 12% in the same field FS_enIMS_ARC.

Its justification names the services it has in mind, and it is worth having exactly as written.

AIML_IMS-MED is the media side of the same idea, and it is a work item rather than a study: it extends TS 26.114 and TS 26.264 so that AI and machine learning data can be carried inside an IMS session AIML_IMS-MED TS 26.114. It has 32 documents [27] and was opened by SP-250878 SP-250878. The reason given is a market one: operators want real-time translation, speech to text, and object detection inside a call [28].

IMSProtoc20-CT is the twentieth-release turn of the protocol alignment work described in Maintenance is the main event: change requests and protocol alignment. It has 19 documents [29], was opened by CP-260044 CP-260044, and its completion field reads 25% IMSProtoc20-CT.

13.5 The row that says two things at once

FS_IMSRE_SEC is the security side of resiliency, and it is the most awkward row in the register. Its own name in the work plan begins with the word Deleted FS_IMSRE_SEC.

At the same time the record holds SP-251237, its study paper, with the outcome word approved against it SP-251237, and 16 meeting documents carrying the acronym FS_IMSRE_SEC. The report it names is TR 33.768, a draft at 0.1.0 TR 33.768.

Both facts are true and this course writes both. The word Deleted is typed into the title by hand — 22 IMS acronyms carry it and one carries STOPPED [30] [31] — and it is the only mark of that kind the work plan has.

It is not a status field, nothing dates it, and nothing says whether it was written before or after the paper was approved.

Its justification, meanwhile, reads like the incident report the Rel-19 study opened with.

13.6 How the newest work reaches the specifications

A work item is only real when it changes a document, so the last check on any of the above is which documents the work names and what state they are in.

Work item The document it writes or changes Where that document stands
FS_IMS_RES TR 29.866 Rel-19 at 19.0.0
IMS_RES-CT TS 23.380 Rel-20 at 20.0.0
SR_IMS TS 26.567 Rel-19 at 19.1.0
FS_IMSResil TR 29.867 Rel-20 at 1.0.0
FS_enIMS_ARC TR 23.801-04 Rel-20 at 0.1.0
AIML_IMS-MED TS 26.114 Rel-20 at 20.0.0
FS_IMSRE_SEC TR 33.768 Rel-20 at 0.1.0

Two things are readable off that table and nothing else is. A version number below 1.0.0 is a working draft, so the architecture study and the security study have no settled text.

And a report that has reached a whole number has stopped being a draft — which is all the number says, and is not a statement that the work behind it is over.

What the table cannot tell you is when any of those versions appeared. The day stored beside each one is the day its file went up on the 3GPP server, and no store here holds a publication day or an approval day for any version.

13.7 Two report numbers that disagree

The Rel-20 architecture study has two document numbers in this data, and they do not agree.

The work item's own table names TR 23.801-03 as the document it will write FS_enIMS_ARC. The specification catalogue says 23.801-03 is a study on mission-critical architecture evolution, owned by SA6, standing at 0.2.0 TR 23.801-03.

Meanwhile 47 IMS documents name TR 23.801-04 in the column that says which document a change targets, its catalogue title is "Study on IMS Architecture Enhancement", its owner is SA2, and it stands at 0.1.0 TR 23.801-04.

Both numbers are in the register. Neither is quietly picked here, because picking one would mean inventing the reason the other is wrong, and no store on this machine holds that reason.

13.8 What the letters between groups show

A liaison is an approved letter from one group to another, and six of them mark this stretch of work.

Four belong to the Release 19 resiliency study, and together they are a conversation. A letter titled "FS_IMS_RES outcome and future work plan" went out from CT4's 125th meeting C4-244485.

Replies came back from CT1's 152nd meeting C1-246921, CT3's 139th C3-250657 and SA2's 166th S2-2413007. One group asked what should follow the study and three answered.

Two more belong to Release 20. CT4 sent "LS on IMS resiliency" from its 134th meeting C4-261402 and CT1 sent the reply from its 161st C1-262608.

The register carries their numbers, titles, the name in the column that says who wrote them, and the meeting. It does not carry a word of what any of them said. A liaison in this record tells you that two groups needed to talk and never what they told each other.

13.9 Reading this chapter in a year's time

Everything above is the state of one store, read once. Three things would change the picture and all three are checkable the same way.

  • R1. New documents under FS_IMSResil or FS_enIMS_ARC. A run of fresh papers is work arriving; silence is only silence in this store, and never proof that anything stopped.

  • R2. TR 29.867 or TR 23.801-04 crossing 1.0.0 TR 29.867 TR 23.801-04. That is the moment a draft stops being a draft.

  • R3. Change requests arriving against TS 23.228 TS 23.228. A study becomes real when the stage 2 document starts changing, and how to count that is The specification family: 241 documents, and the two that are IMS.

How to run those three checks, and the full list of what this record cannot tell you whatever you check, is Reading the record yourself, and what it cannot tell you.

Where the numbers in this chapter come from

  1. why FS_IMS_RES was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1010003/CP-232281-CP-232026/CP-232281-CP-232026.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  2. what FS_IMS_RES set out to do the opening sentences of the objective of the work item description /var/www/whatthespec.net/data/data/wis/1010003/CP-232281-CP-232026/CP-232281-CP-232026.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  3. what IMS_RES-CT set out to do the opening sentences of the objective of the work item description /var/www/whatthespec.net/data/data/wis/1080001/CP-253168/CP-253168.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  4. what SR_IMS set out to do the opening sentences of the objective of the work item description /var/www/whatthespec.net/data/data/wis/1030003/SP-240492 - revision of SP-240061 New WID on Split Rendering over IMS (SR_IMS)/SP-240492 - revision of SP-240061 New WID on Split Rendering over IMS (SR_IMS).md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  5. why FS_IMSResil was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1080002/CP-251280 was CP-251017/CP-251280 was CP-251017.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  6. what FS_enIMS_ARC set out to do the opening sentences of the objective of the work item description /var/www/whatthespec.net/data/data/wis/1100026/SP-251692 was SP-251656 was of SP-251638 was SP-251411_S2-2511300_was_S2-2511257_New SID on IMS architecture enhancement r02/SP-251692 was SP-251656 was of SP-251638 was SP-251411_S2-2511300_was_S2-2511257_New SID on IMS architecture enhancement r02.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  7. why FS_enIMS_ARC was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1100026/SP-251692 was SP-251656 was of SP-251638 was SP-251411_S2-2511300_was_S2-2511257_New SID on IMS architecture enhancement r02/SP-251692 was SP-251656 was of SP-251638 was SP-251411_S2-2511300_was_S2-2511257_New SID on IMS architecture enhancement r02.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  8. why FS_IMSRE_SEC was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1090015/SP-251237_new_SID_Security Aspects for IMS resiliency/SP-251237_new_SID_Security Aspects for IMS resiliency.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  9. the newest IMS document on record went up on 2026-08-04 the largest upload stamp over the IMS documents, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  10. 4 IMS work plan rows are marked Rel-18 rows of table `item` in /var/www/whatthespec.net/data/database/api/wp.sqlite taken in by the acronym rule whose release field is exactly 'Rel-18', read 2026-08-05
  11. 217 IMS documents are marked Rel-18 IMS documents whose release column is 'Rel-18', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. That is the release the document was filed against and nothing more
  12. 10 IMS work plan rows are marked Rel-19 rows of table `item` in /var/www/whatthespec.net/data/database/api/wp.sqlite taken in by the acronym rule whose release field is exactly 'Rel-19', read 2026-08-05
  13. 387 IMS documents are marked Rel-19 IMS documents whose release column is 'Rel-19', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. That is the release the document was filed against and nothing more
  14. 7 IMS work plan rows are marked Rel-20 rows of table `item` in /var/www/whatthespec.net/data/database/api/wp.sqlite taken in by the acronym rule whose release field is exactly 'Rel-20', read 2026-08-05
  15. 189 IMS documents are marked Rel-20 IMS documents whose release column is 'Rel-20', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. That is the release the document was filed against and nothing more
  16. 45 documents carry the work item IMSProtoc18 rows of `tdoc` whose work item column names IMSProtoc18 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  17. 209 IMS documents were filed in 2024 IMS documents whose upload stamp starts 2024, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  18. 365 IMS documents were filed in 2025 IMS documents whose upload stamp starts 2025, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  19. 166 IMS documents were filed in 2026 IMS documents whose upload stamp starts 2026, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  20. 237 IMS documents carry no upload time IMS documents whose upload column is empty, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05, which is why the per-year counts do not add up to the total and why a count over a year and a count over the whole record answer different questions
  21. 147 documents carry the work item FS_IMS_RES rows of `tdoc` whose work item column names FS_IMS_RES as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  22. 101 documents carry the work item IMS_RES-CT rows of `tdoc` whose work item column names IMS_RES-CT as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  23. 84 documents carry the work item SR_IMS rows of `tdoc` whose work item column names SR_IMS as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  24. why SR_IMS was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1030003/SP-240492 - revision of SP-240061 New WID on Split Rendering over IMS (SR_IMS)/SP-240492 - revision of SP-240061 New WID on Split Rendering over IMS (SR_IMS).md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  25. 93 documents carry the work item FS_IMSResil rows of `tdoc` whose work item column names FS_IMSResil as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  26. 51 documents carry the work item FS_enIMS_ARC rows of `tdoc` whose work item column names FS_enIMS_ARC as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  27. 32 documents carry the work item AIML_IMS-MED rows of `tdoc` whose work item column names AIML_IMS-MED as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  28. why AIML_IMS-MED was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1080049/SP-250878_SP-250634/SP-250878_SP-250634.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  29. 19 documents carry the work item IMSProtoc20-CT rows of `tdoc` whose work item column names IMSProtoc20-CT as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  30. 22 IMS work item acronyms carry the word Deleted at the front of their own name rows taken in by the acronym rule whose name field begins 'Deleted', read 2026-08-05. The prefix is written by hand into the title and is the only mark of the kind the work plan carries; it is not a status field (FS_IMSRE_SEC, FS_eIMS_IBCF, IMS-EMER, IMS-RT, IMS-SE, IMS8, IMSLocal, IMSProtoc5, IMSProtoc6, IMS_Comm_GqRx_Harm, IMS_RegCon, IMS_RegCon-CT, IMS_RegCon-CT1, IMS_RegCon-CT4, IMS_RegCon-del1, IMS_RegCon-del2, IMS_TELEP_S4, IMS_WebRTC-holmb, IMS_WebRTC-yusef, IRTSD-IMSCS, RANimp-RABSE-IuEnhIMS, VoIMS)
  31. 1 IMS work item acronyms carry the word STOPPED at the front of their own name rows taken in by the acronym rule whose name field begins 'STOPPED', read 2026-08-05. The prefix is written by hand into the title and is the only mark of the kind the work plan carries; it is not a status field (IMSProtoc17_dataCH-UEContest)
  32. 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

Every source this course is built on

Check yourself

Answers appear when you pick one, with where they come from.

Q13.1 Why was the Release 19 resiliency study opened?

Q13.2 What does the Release 20 architecture study give as its reason for existing?

Q13.3 The Release 20 architecture study names one report number and 47 IMS documents name another. What does this course do?

Q13.4 One Release 20 study carries "Deleted" at the front of its own name. What else does the record hold about it?

Q13.5 How many IMS work plan rows are marked Release 20?

This chapter was built from a source register generated 2026-08-05. A fresher build of the register may hold different numbers.