FS NC8200-4TD: the 4-slot chassis as a stated 4x NC8200-24BC fill (96x 25G SFP28 + 8x 100G QSFP28) #1690

Merged
jaylouisw merged 2 commits from catalogue/1116-1626 into main 2026-11-16 16:33:41 +00:00
Owner

Fixes #760

The issue asks for the NC8200-4TD and lists its line cards (NC8200-24BC, NC8200-8C). The frame
ships unloaded and a slot is a bay, not a port, so what is modelled is a stated complement -
the convention docs/MAINTAINING.md records for devices/arista/7504r.json ("7504R (4x 36CQ)"):
the card complement goes in the label.

What was added

devices/fs/nc8200-4td.json - NC8200-4TD (4x NC8200-24BC), 2U, 4 line-card slots:

claim ports modelled
96x 25G SFP28 (TFGigabitEthernet<slot>/1-24) 96, connector: sfp28, speedCaps [25000, 10000]
8x 100G QSFP28 (HundredGigabitEthernet<slot>/1-2) 8, connector: qsfp28, speedCaps [100000, 40000, 25000, 10000]
RJ-45 RS-232 console 1, type: console

Capabilities l2-forward, l3-forward, cli; 105 ports total.

Sources

Datasheet vs product page, stated not averaged

The datasheet says 6.4 Tbps switching capacity and 4.76 Bpps; FS's own product page prints
3.2 Tbps for the same chassis. 6.4 Tbps is the duplex aggregate - 3.2 Tbps each way - and 96x 25G

  • 8x 100G is exactly 3.2 Tbps per direction, so a full 24BC load is line-rate and not oversubscribed.
    fabric.capacityMbps is therefore 3,200,000 (the per-direction figure the fabric audit compares
    against), with both vendor numbers recorded in the note and in the fabric note.

Not modelled, and why

  • The other two fills: 4x NC8200-8C = 32x 100G and 4x NC8200-16Q = 64x 40G. Different boxes, their
    own specs - model is unique per vendor, so they cannot be variants of this file.
  • The "128x 10G/25G" headline density is this fill with the eight QSFP28 cages broken out, not a
    fourth card type; it is in the note.
  • One out-of-band RJ45 management port and one Mini-USB B console port are listed without a stated
    rate (management) or as the same console in another connector (Mini-USB B), so neither is modelled.
  • No PoE anywhere in the datasheet: 2x 1+1 hot-swap AC, 600 W max consumption, no PSE, no budget.

Gates

node scripts/syntax-check.js          -> Syntax-checked 354 files: all OK
node scripts/audit-port-types.js      -> No port contradicts its declared type. (exit 0)
node test/hardware-invariants.test.js -> 26 passed, 0 failed
node test/spec-vs-description.test.js -> 3 passed, 0 failed
node test/device-aliases.test.js      -> 13 passed, 0 failed
importer roundtrip (PF_INCOMING/PF_DEVICES on a scratch copy):
  Scanned 1 file(s), 1 spec(s) / Added: 1  duplicates: 0  rejected: 0  (no capability warnings)
  the re-imported file is identical to devices/fs/nc8200-4td.json
fabric audit: offered 3,200,000 / capacity 3,200,000, ratio 1.0, exceeds false
npm run regen:registry-doc            -> 2,239 models, 25,806 sockets (README, CONNECTORS, DEVICE_REGISTRY, wiki)

Every number above appears in one of the three FS sources; nothing in this PR is unverifiable.

Fixes #760 The issue asks for the NC8200-4TD and lists its line cards (NC8200-24BC, NC8200-8C). The frame ships unloaded and a slot is a bay, not a port, so what is modelled is a **stated complement** - the convention `docs/MAINTAINING.md` records for `devices/arista/7504r.json` ("7504R (4x 36CQ)"): the card complement goes in the label. ## What was added `devices/fs/nc8200-4td.json` - NC8200-4TD (4x NC8200-24BC), 2U, 4 line-card slots: | claim | ports modelled | |---|---| | 96x 25G SFP28 (`TFGigabitEthernet<slot>/1-24`) | 96, `connector: sfp28`, `speedCaps [25000, 10000]` | | 8x 100G QSFP28 (`HundredGigabitEthernet<slot>/1-2`) | 8, `connector: qsfp28`, `speedCaps [100000, 40000, 25000, 10000]` | | RJ-45 RS-232 console | 1, `type: console` | Capabilities `l2-forward, l3-forward, cli`; 105 ports total. ## Sources - Datasheet (text layer read with pdftotext, not a rendered summary): <https://resource.fs.com/mall/doc/20221114163118tn7bta.pdf> - Table 1 (4 line-card slots, max densities), Table 2 (per-card port density and the QSFP28 breakout note), Table 5 (switching capacity, forwarding rate), power/fan table, Table 4 (stacking). - Product page <https://www.fs.com/products/110482.html> - line-card set and the "96x 25G with 8x 100G, 64x 40G, or 32x 100G" configuration sentence. - FSOS CLI reference for N5860/N8560/NC8200 <https://resource.fs.com/mall/doc/20230602150328dcg5nv.pdf> - interface naming (`TFGigabitEthernet`, `HundredGigabitEthernet`, `<slot>/<port>`, line-card slot numbering as in `switch-mode ... slot 3`). ## Datasheet vs product page, stated not averaged The datasheet says **6.4 Tbps** switching capacity and **4.76 Bpps**; FS's own product page prints **3.2 Tbps** for the same chassis. 6.4 Tbps is the duplex aggregate - 3.2 Tbps each way - and 96x 25G + 8x 100G is exactly 3.2 Tbps per direction, so a full 24BC load is line-rate and not oversubscribed. `fabric.capacityMbps` is therefore **3,200,000** (the per-direction figure the fabric audit compares against), with both vendor numbers recorded in the note and in the fabric note. ## Not modelled, and why - The other two fills: 4x NC8200-8C = 32x 100G and 4x NC8200-16Q = 64x 40G. Different boxes, their own specs - `model` is unique per vendor, so they cannot be variants of this file. - The "128x 10G/25G" headline density is this fill with the eight QSFP28 cages broken out, not a fourth card type; it is in the note. - One out-of-band RJ45 management port and one Mini-USB B console port are listed without a stated rate (management) or as the same console in another connector (Mini-USB B), so neither is modelled. - No PoE anywhere in the datasheet: 2x 1+1 hot-swap AC, 600 W max consumption, no PSE, no budget. ## Gates ``` node scripts/syntax-check.js -> Syntax-checked 354 files: all OK node scripts/audit-port-types.js -> No port contradicts its declared type. (exit 0) node test/hardware-invariants.test.js -> 26 passed, 0 failed node test/spec-vs-description.test.js -> 3 passed, 0 failed node test/device-aliases.test.js -> 13 passed, 0 failed importer roundtrip (PF_INCOMING/PF_DEVICES on a scratch copy): Scanned 1 file(s), 1 spec(s) / Added: 1 duplicates: 0 rejected: 0 (no capability warnings) the re-imported file is identical to devices/fs/nc8200-4td.json fabric audit: offered 3,200,000 / capacity 3,200,000, ratio 1.0, exceeds false npm run regen:registry-doc -> 2,239 models, 25,806 sockets (README, CONNECTORS, DEVICE_REGISTRY, wiki) ``` Every number above appears in one of the three FS sources; nothing in this PR is unverifiable.
FS NC8200-4TD: the 4-slot chassis as a stated 4x NC8200-24BC fill, 96x 25G SFP28 + 8x 100G QSFP28 (#760)
All checks were successful
CI / Classify the change (pull_request) Successful in 13s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 17s
CI / Server smoke test (pull_request) Successful in 24s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Docker build (pull_request) Successful in 32s
CI / Test (Node 20) (pull_request) Successful in 1m19s
CI / Test (Node 22) (pull_request) Successful in 1m25s
48fffeee15
The request lists the chassis unloaded with its line cards. A 4-slot frame has no ports of
its own, so the spec is the stated complement: 4x NC8200-24BC gives 96x 25G SFP28 and
8x 100G QSFP28, which is the datasheet-era "96x 25G with 8x 100G" headline. The 32x 100G
(4x NC8200-8C) and 64x 40G (4x NC8200-16Q) fills are different boxes and are left for
their own specs.

Fabric: the datasheet (Table 5) states 6.4 Tbps / 4.76 Bpps and product page 110482 states
3.2 Tbps for the same chassis. 6.4 is the duplex aggregate, 3.2 per direction; the port
census sums to exactly 3.2 Tbps per direction, so fabric.capacityMbps is 3,200,000 and the
full 24BC load is line-rate. The discrepancy is stated in the note rather than averaged.
jaylouisw scheduled this pull request to auto merge when all checks succeed 2026-11-16 16:33:31 +00:00
jaylouisw force-pushed catalogue/1116-1626 from 48fffeee15
All checks were successful
CI / Classify the change (pull_request) Successful in 13s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 17s
CI / Server smoke test (pull_request) Successful in 24s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Docker build (pull_request) Successful in 32s
CI / Test (Node 20) (pull_request) Successful in 1m19s
CI / Test (Node 22) (pull_request) Successful in 1m25s
to f3d53d4bfa
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 13s
CI / Test (Node 18) (pull_request) Has been skipped
CI / Test (Node 20) (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Has been skipped
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
2026-11-16 16:33:34 +00:00
Compare
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!1690
No description provided.