Juniper EX2300: four more models, a 740 W budget one Juniper guide calls 750 W, and two capacity figures Juniper disagrees with itself on #1564

Merged
jaylouisw merged 2 commits from catalogue/1115-2326 into main 2026-11-16 01:36:00 +00:00
Owner

What

The Juniper EX2300 access-switch family: six models now carry their real port layout, PoE budget and cage, up from two thin entries and a portless stub.

model access ports PoE uplinks capacity
EX2300-24T (corrected) 24x 1G RJ-45 none 4x 1G/10G SFP/SFP+ 128 Gbps
EX2300-48T (corrected) 48x 1G RJ-45 none 4x SFP/SFP+ 176 Gbps
EX2300-24P (new) 24x 1G RJ-45 24 ports, 370 W 4x SFP/SFP+ 128 Gbps
EX2300-48P (new) 48x 1G RJ-45 48 ports, 740 W 4x SFP/SFP+ 176 Gbps
EX2300-24MP (new) 16x 1G + 8x 1/2.5G RJ-45 24 ports, 380 W 4x SFP/SFP+ 128 Gbps (see below)
EX2300-48MP (new) 32x 1G + 16x mGig RJ-45 48 ports, 740 W 6x SFP/SFP+ 176 Gbps (see below)

Also removed devices/juniper/ex2300.json: model EX2300 is not a SKU, it carried the label EX2300-24T (so a correct 24T spec collided with it by label) and it declared no ports at all.

Two vendor documents contradict each other, and both are recorded

  1. PoE budget of the 48P and the 48MP. The datasheet, the power-supply table and the system-overview table all say 740 W; the PoE User Guide, Table 5, says 750 W for the same two models. 740 W is modelled, because 48 x 15.4 W = 739.2 W lands exactly on it and the datasheet spells that arithmetic out.
  2. Switching capacity of the multigigabit models. The datasheet groups the 24MP with the 24-port models (128 Gbps) and the 48MP with the 48-port models (176 Gbps). Juniper's per-model hardware specs give the 24MP 152 Gbps bidirectional and the 48MP 264 Gbps — a different number for the same box in each case. The datasheet figure is what is modelled; the conflict is stated in each spec's fabric.note.

Modelling decisions worth a reviewer's eye

  • The MP models' port numbering is not contiguous by rate (24MP: mGig 0-7, 1G 8-23; 48MP: 1G 0-15, mGig 16-31, 1G 32-47), so they are listed explicitly instead of from one template.
  • The 48MP's 16 multi-gigabit ports are listed by Juniper as 100M/1G/2.5G/5G/10G. They are modelled at 2.5G with caps to 5G: engine/media.js marks 10GBASE-T as carrying no power, and every one of those ports is a PSE port, so a 10G model speed would silently strip PoE. The 5G/10G rates are recorded in notes.
  • 802.3af/at only — no EX2300 does 802.3bt, including the multigigabit models.
  • Console is not declared as an interface (per existing convention): RJ-45 serial, plus mini-USB Type B on the four non-MP models.

Drive-by: a red suite on main

test/spec-vs-description.test.js was failing on origin/main before this branch: the four Edgecore AS4610 descriptions from #1555 claimed "2x 20G QSFP+" stacking cages that the same PR deliberately did not model as ports, and the test counts that claim. Reworded to "two 20G QSFP+ …" in the four specs and in research/incoming/devices-edgecore-as4610.json; no port data changed.

Sources (every field carries one)

Validation

  • node scripts/syntax-check.js — 327 files OK
  • node scripts/audit-port-types.js — 2081 models / 21845 ports, 0 offending devices, 0 offending ports
  • importer: --apply --upgrade — 4 added, 2 upgraded (EX2300-24T and -48T detail 0 -> 10), capability tripwire silent apart from the expected 802.3af/802.3at "kept as declared" notes, which the engine does not implement
  • npm run regen:registry-doc — 2,078 -> 2,081 models, switch 310 -> 313
  • npm test — 257 suites, 9371 assertions, 0 failed

Not verified

  • The EX2300-C (fanless 12-port) is a separate datasheet and was left out of this increment.
  • Juniper's console default line rate is not stated in any page fetched here, so no baud rate is claimed.
## What The Juniper EX2300 access-switch family: **six models** now carry their real port layout, PoE budget and cage, up from two thin entries and a portless stub. | model | access ports | PoE | uplinks | capacity | |---|---|---|---|---| | EX2300-24T (corrected) | 24x 1G RJ-45 | none | 4x 1G/10G SFP/SFP+ | 128 Gbps | | EX2300-48T (corrected) | 48x 1G RJ-45 | none | 4x SFP/SFP+ | 176 Gbps | | EX2300-24P (new) | 24x 1G RJ-45 | 24 ports, 370 W | 4x SFP/SFP+ | 128 Gbps | | EX2300-48P (new) | 48x 1G RJ-45 | 48 ports, 740 W | 4x SFP/SFP+ | 176 Gbps | | EX2300-24MP (new) | 16x 1G + 8x 1/2.5G RJ-45 | 24 ports, 380 W | 4x SFP/SFP+ | 128 Gbps (see below) | | EX2300-48MP (new) | 32x 1G + 16x mGig RJ-45 | 48 ports, 740 W | 6x SFP/SFP+ | 176 Gbps (see below) | Also **removed** `devices/juniper/ex2300.json`: model `EX2300` is not a SKU, it carried the label `EX2300-24T` (so a correct 24T spec collided with it by label) and it declared no ports at all. ## Two vendor documents contradict each other, and both are recorded 1. **PoE budget of the 48P and the 48MP.** The datasheet, the power-supply table and the system-overview table all say **740 W**; the *PoE User Guide*, Table 5, says **750 W** for the same two models. 740 W is modelled, because 48 x 15.4 W = 739.2 W lands exactly on it and the datasheet spells that arithmetic out. 2. **Switching capacity of the multigigabit models.** The datasheet groups the 24MP with the 24-port models (**128 Gbps**) and the 48MP with the 48-port models (**176 Gbps**). Juniper's per-model hardware specs give the 24MP **152 Gbps** bidirectional and the 48MP **264 Gbps** — a different number for the same box in each case. The datasheet figure is what is modelled; the conflict is stated in each spec's `fabric.note`. ## Modelling decisions worth a reviewer's eye - The MP models' port numbering is not contiguous by rate (24MP: mGig 0-7, 1G 8-23; 48MP: 1G 0-15, mGig 16-31, 1G 32-47), so they are listed explicitly instead of from one template. - The 48MP's 16 multi-gigabit ports are listed by Juniper as 100M/1G/2.5G/**5G/10G**. They are modelled at 2.5G with caps to 5G: `engine/media.js` marks 10GBASE-T as carrying no power, and every one of those ports is a PSE port, so a 10G model speed would silently strip PoE. The 5G/10G rates are recorded in `notes`. - 802.3af/at only — no EX2300 does 802.3bt, including the multigigabit models. - Console is not declared as an interface (per existing convention): RJ-45 serial, plus mini-USB Type B on the four non-MP models. ## Drive-by: a red suite on main `test/spec-vs-description.test.js` was failing on `origin/main` before this branch: the four Edgecore AS4610 descriptions from #1555 claimed **"2x 20G QSFP+"** stacking cages that the same PR deliberately did not model as ports, and the test counts that claim. Reworded to "two 20G QSFP+ …" in the four specs and in `research/incoming/devices-edgecore-as4610.json`; no port data changed. ## Sources (every field carries one) - per-model hardware specs: <https://apps.juniper.net/hct/product/EX2300/hwspecs>, <https://apps.juniper.net/hct/product/EX2300%20Multigigabit/hwspecs> - power and PoE budget: <https://www.juniper.net/documentation/us/en/hardware/ex2300/topics/topic-map/ex2300-power-supplies.html>, <https://www.juniper.net/documentation//us/en/hardware/ex2300/topics/topic-map/ex2300-system-overview.html> - firmware / first Junos release, port roles: <https://www.juniper.net/documentation/us/en/hardware/ex2300/ex2300.pdf> - 48MP multi-gigabit rates: <https://www.juniper.net/documentation/us/en/hardware/qsg/ex2300/ex2300.pdf> - the 750 W contradiction: <https://www.juniper.net/documentation/us/en/software/junos/poe/poe.pdf> - switching capacity, licence tiers, L2/L3 scale: EX2300 datasheet 1000579-018-EN (Apr 2019), read from a Juniper-hosted mirror at <https://juniper.vn/public/cuploads/files/datasheet-juniper-ex2300.pdf> — Juniper's own `juniper.net/.../ex2300-line-of-ethernet-switches-datasheet.pdf` path now 404s. ## Validation - `node scripts/syntax-check.js` — 327 files OK - `node scripts/audit-port-types.js` — 2081 models / 21845 ports, **0 offending devices, 0 offending ports** - importer: `--apply --upgrade` — 4 added, 2 upgraded (`EX2300-24T` and `-48T` detail 0 -> 10), capability tripwire silent apart from the expected `802.3af`/`802.3at` "kept as declared" notes, which the engine does not implement - `npm run regen:registry-doc` — 2,078 -> 2,081 models, switch 310 -> 313 - `npm test` — **257 suites, 9371 assertions, 0 failed** ## Not verified - The EX2300-C (fanless 12-port) is a separate datasheet and was left out of this increment. - Juniper's console default line rate is not stated in any page fetched here, so no baud rate is claimed.
Juniper EX2300: four more models, a 740 W budget one Juniper guide calls 750 W, and two capacity figures Juniper disagrees with itself on
All checks were successful
CI / Classify the change (pull_request) Successful in 14s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 18s
CI / Test (Node 18) (pull_request) Successful in 3m28s
CI / Test (Node 20) (pull_request) Successful in 3m11s
CI / Server smoke test (pull_request) Successful in 27s
CI / Docker build (pull_request) Successful in 23s
CI / Test (Node 22) (pull_request) Successful in 3m4s
b781391993
The EX2300 family was in the catalogue as two thin entries (24T, 48T) plus a
portless stub; the eight-SKU range is now six models with their real port
layout, PoE budget and cage.

New: EX2300-24P, EX2300-48P, EX2300-24MP, EX2300-48MP.
Corrected: EX2300-24T and EX2300-48T (no connector on the access ports, no cage
on the uplinks, no PoE statement, no source).
Removed: devices/juniper/ex2300.json - model "EX2300" is not a SKU, it carried
the label "EX2300-24T", and it had no ports at all.

Two vendor documents contradict each other and both are recorded rather than
smoothed over:
  - PoE budget for the 48P and 48MP: 740 W in the datasheet, the power-supply
    table and the system-overview table; 750 W in PoE User Guide Table 5.
    740 W is modelled, because 48 x 15.4 W = 739.2 W.
  - Switching capacity for the multigigabit models: the datasheet groups the
    24MP with 128 Gbps and the 48MP with 176 Gbps, while the per-model hardware
    specs give 152 and 264 Gbps. The datasheet figure is modelled and the
    conflict is named in each spec's fabric note.

Also fixes the four Edgecore AS4610 descriptions from #1555, which claimed
"2x 20G QSFP+" for stacking cages the same PR deliberately did not model -
test/spec-vs-description.test.js counted them and the suite was red on main.
jaylouisw force-pushed catalogue/1115-2326 from b781391993
All checks were successful
CI / Classify the change (pull_request) Successful in 14s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 18s
CI / Test (Node 18) (pull_request) Successful in 3m28s
CI / Test (Node 20) (pull_request) Successful in 3m11s
CI / Server smoke test (pull_request) Successful in 27s
CI / Docker build (pull_request) Successful in 23s
CI / Test (Node 22) (pull_request) Successful in 3m4s
to 1d6030a589
Some checks failed
CI / Classify the change (pull_request) Successful in 14s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 19s
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Successful in 3m34s
CI / Test (Node 18) (pull_request) Successful in 3m55s
CI / Test (Node 20) (pull_request) Failing after 12m19s
2026-11-16 01:11:57 +00:00
Compare
Author
Owner

This branch conflicts with main, so it was not rebased:

devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json 

A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.

This branch conflicts with `main`, so it was not rebased: ``` devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json ``` A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.
jaylouisw force-pushed catalogue/1115-2326 from 1d6030a589
Some checks failed
CI / Classify the change (pull_request) Successful in 14s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 19s
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Successful in 3m34s
CI / Test (Node 18) (pull_request) Successful in 3m55s
CI / Test (Node 20) (pull_request) Failing after 12m19s
to d8a9229dc2
Some checks failed
CI / Classify the change (pull_request) Successful in 18s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 28s
CI / Test (Node 20) (pull_request) Failing after 28s
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Successful in 3m34s
CI / Test (Node 18) (pull_request) Successful in 3m56s
2026-11-16 01:34:36 +00:00
Compare
jaylouisw force-pushed catalogue/1115-2326 from d8a9229dc2
Some checks failed
CI / Classify the change (pull_request) Successful in 18s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 28s
CI / Test (Node 20) (pull_request) Failing after 28s
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Successful in 3m34s
CI / Test (Node 18) (pull_request) Successful in 3m56s
to e74a7f5689
Some checks failed
CI / Classify the change (pull_request) Successful in 24s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 19s
CI / Server smoke test (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Successful in 3m23s
CI / Test (Node 18) (pull_request) Successful in 3m51s
CI / Docker build (pull_request) Failing after 13m3s
CI / Test (Node 20) (pull_request) Failing after 13m39s
2026-11-16 01:35:43 +00:00
Compare
Author
Owner

This branch conflicts with main, so it was not rebased:

devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json 

A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.

This branch conflicts with `main`, so it was not rebased: ``` devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json ``` A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.
Author
Owner

This branch conflicts with main, so it was not rebased:

devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json 

A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.

This branch conflicts with `main`, so it was not rebased: ``` devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json ``` A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.
Author
Owner

This branch conflicts with main, so it was not rebased:

devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json 

A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.

This branch conflicts with `main`, so it was not rebased: ``` devices/edgecore/as4610-30p.json devices/edgecore/as4610-30t.json devices/edgecore/as4610-54p.json devices/edgecore/as4610-54t.json devices/juniper/ex2300-24mp.json devices/juniper/ex2300-24p.json devices/juniper/ex2300-24t.json devices/juniper/ex2300-48mp.json devices/juniper/ex2300-48p.json devices/juniper/ex2300-48t.json devices/juniper/ex2300.json docs/DEVICE_REGISTRY.md research/incoming/devices-edgecore-as4610.json research/incoming/devices-juniper-ex2300.json ``` A conflict is a decision about behaviour rather than a mechanical fix, so this is left for a human. Rebasing by hand and force-pushing will let the automation pick it up again.
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge!1564
No description provided.