[Device]: OptiXtrans E6616X #803
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
jaylouisw/packetforge#803
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Vendor
Huawei
Exact model name
OptiXtrans E6616X
Category
Other / not sure
Ports and interfaces
PoE budget (watts)
None (no PoE output)
Datasheet or product page URL
https://e.huawei.com/en/products/optical-transmission/optixtrans-e6600
What do you want to model with it?
Metro aggregation MS-OTN node (grey/backplane-focused variant).
Capabilities
MS-OTN; OTN/PKT/SDH unified switching
Before submitting
Closing status: blocked on a stated configuration.
The OptiXtrans E6600/E9600 specification tables publish what a frame can do - service board slots (E6600: "Number of Service Board Slots: DC chassis 14 / AC chassis 12"), cross-connect capacity, maximum wavelengths and per-channel rate - and nothing about ports, because an OTN subrack has none until service boards are fitted. A 400G/λ, 96-wavelength figure is a capacity, not a socket count.
So the issue stays open;
docs/DEVICE_REGISTRY.md§"What is left open, and why" records the reason for the family (#803, #804, #807, #808, #810).What would close it: the board set you intend to model (e.g. "E6616 with two TN13LSC boards"), or a decision to model the frame's slots as a repeatable module - which would be an engine feature, not a spec file.
Re-checked against the model-specific brochure this pass; the blocker is narrower and now exact.
Sources read (all vendor):
OptiXtrans E6616X V100R026C00 Product Brochure (26H2), read end to end (128 paragraphs, 3 tables, 6 images).What the brochure does publish: 780G OTN cross-connect (TM65UXCL needs a licence to reach it), 620G OSU, 160G VC-4 / 20G VC-3/VC-12 SDH, 820G packet, 14/12 service slots, line rate 2.5/10/100 Gbit/s gray light, pluggables SFP/eSFP, XFP/TXFP, SFP+/TSFP+, QSFP+, QSFP28, SFP28, MTBF 111.9 years.
What it does not publish: any service-board table. The document carries zero board part numbers (no B1L, TMB, TMK or F3 order codes anywhere), so "14 slots" has no port census behind it. The only board it names, TM65UXCL, is the system-control/cross-connect board, not a service board. A 14-slot frame with no stated card has no cableable port, and
loadCatalog()refuses a spec with none, so #803 stays open rather than carrying an invented panel.One correction worth recording: the EN spec table's merged cells misplace the "100 Gbit/s (grey)" value. The FR table and the brochure both put 400 Gbit/s on E6616 / E6608 / E6608T and 100 Gbit/s grey on E6616X, which agrees with the E6616X brochure's "Maximum single-channel rate 100 Gbit/s (gray light)" and with its N/A wavelength rows: the X has no WDM line.
Exact missing fields: (1) the service-board set valid in an E6616X subrack, with ports per board; (2) whether the fgOTN/OSU client ports are QSFP28-only at 100G.
What would close it: a stated configuration of the shape already used for E6608/E6616 ("E6616X DC chassis with 14x "), or Huawei's board table for the V100R026 hardware-description generation.
Re-checked. Still blocked on the service-board census, but the frame side is now vendor-sourced from a downloadable chapter, and one trap in this thread is closed. PR #1826 records it (
research/optixtrans-e6600-chassis-chapters.md); it does not close the issue.New route (no login).
https://e.huawei.com/en/documents/others/<32-hex>material pages are server-rendered and print the real asset path in their HTML, so only the material hash has to be found. That fetched OptiXtrans E6600 V100R025C10 Product Documentation, chapter 1 (E6616 Chassis), 20 pp, 1,535,511 B, issue 2026-03-31.Slot layout, verbatim: DC chassis service slots 1-14, system control 15/16, DC power 17-18, fan 19. AC chassis service slots 2, 4, 5-14, printed as 12/10 - 1500 W AC chassis (dual power supply): 12; 2400 W AC chassis (four power supply): 10 - AC power 1 and 3 or 1-4 for
TMB1APIU.Correction to what this thread assumes: the spec table row "AC chassis: 12" is one of two AC variants. A 2400 W AC chassis has 10 service slots, not 12, because four slots carry APIU boards. A spec label that omits the power configuration is off by two.
Still missing: the service-board part number valid in the E6616X, plus that board front-panel census (count, rate, cage). The chassis chapter names only system-control boards (
TMK5SXCH,TMK5XCH,TMK6XCH,TMK5UXCME,TMK5XCS), fgOTN cards (TMK1FXCL,TMK1XCL),TMB1APIUandTMK5FAN; section 1.5 is per-slot backplane capacity, not ports. The E6600 board mirror covers E6608T/E6608/E6616 - not the X.What would close it now: any
TMB1*/TMK1*chapter whose Valid Slots table names OptiXtrans E6616X (fetch it withscripts/huawei-material-fetch.js <hash> <dir>), a readable EDOC1100456166 section, or a stated configuration.Re-checked. One premise in this record was wrong and is now corrected - but the issue is still blocked, on the client board chapter, not on the hedex wall. PR #1871 carries the evidence (
research/optixtrans-e-file-numbered-bundle.md).The correction. "The E6600 generation publishes its board list only inside the closed hedex tree" is false as stated. E6600 documentation is published one chapter per PDF on
e-file.huawei.com, anonymously, and two board chapters were fetched and read end to end today:(6)-UXCME— https://e-file.huawei.com/marketingcloud/pep/asset/20000001/Material/5fa3c10e4f454dfaa44f26447563718d/M3T1A669N1304588494599672043/_6_-UXCME.pdf — 200,application/pdf, 590,457 B;TMK5UXCMEsystem control/switching/timing, front panel, indicators, micro switches.(21)-TMB3AUX— https://e-file.huawei.com/marketingcloud/pep/asset/20000001/Material/104eb87894f1489dac28e58171dca5cc/M3T1A669N1304590990909157504/_21_-TMB3AUX.pdf — 200,application/pdf, 482,177 B; AUX board, dry contact + orderwire, clock cascading onCLK1/TOD1-CLK3/TOD3orHP1....The route is now exhausted, measured. The "Related Materials" list on these pages is one uploader's bundle, not a chapter index: crawling it breadth-first from the five optical pages on this family's threads drains at 40 pages, 27 of them numbered
(2)-(28)with only(1)unpublished. It contains exactly two board chapters - the two above - and no client board. Numbering cannot be extended by arithmetic.Why this still does not close #803, and one new obstacle. Neither chapter carries a
Valid Slotstable (Valid Slotoccurs 0 times in both) and neither names any subrack (E6616X0,E6608T0,E66160,E96050). That is a template difference from the E9600 side, whereTNG2DAP§1.9 does publish a board-to-subrack matrix — so an E6600 client-board chapter, when one surfaces, may still not state that the board is valid in the X.Also re-tested today (so it is not re-walked): hedex is closed for every user-agent —
support.huawei.com/hedex/hdx.do?docid=EDOC1100163901403,info.support.huawei.com/hedex/api/doc/EDOC1100163901and/hedex/api/pages/302, and even the Doradohedex/api/pages/EDOC1100214756/...path that answered 200 this morning now answers 302. Wayback CDX returns[]forsupport.huawei.com/hedex/pages/EDOC1100156575*andhdx.do?docid=EDOC1100156575*(theinfo.host answered IA 503, so unverified, not absent). Andhttps://e.huawei.com/en/documents/others/dc67ba7b606b46df975107406dafb0a1, which search still lists as the E6616 V100R025C10 brochure, now serves a 404 page with no asset path.Exact missing field, unchanged: the service-board part number valid in an E6616X subrack, plus that board's front-panel port census.
What would close it: a stated configuration line —
E6616X DC (12x <board>)— from an E6600 ordering/configuration guide, or one client-board chapter from the E6600 V100R025C10/C10 documentation set together with a statement of X validity.New route found and drained in the same pass: Huawei's 26H2 brochure tree hands over its PDFs, and neither E6616X brochure carries a client board. Still blocked on the board census; PR #1879 is catalogue hygiene for this family, not a close.
The route, which is new to this thread.
https://e.huawei.com/en/documents/products/enterprise-optical-network/<32-hex>is server-rendered like thedocuments/otherspages and prints the asset path inline, so it needs no asset token. Breadth-first from the E6616 brochure page (dc67ba7b606b46df975107406dafb0a1) it exposes seven product brochures: E6616, E6616X, E6608, E6608T, OSN 1800 V, OSN 1800 V Pro, OSN 1800 I Enhanced — all "26H2" / V100R026C00.What the E6616X ones say. Both are HTTP 200 with a search-crawler user-agent (a desktop UA gets 403 Access Denied, and so does
e-file.huawei.comon the same asset path):.../Material/91ebb26076e6453ab0c72480cd6e5464/M3T1A590N1296415560500404267/OptiXtrans E6616X V100R026C00 Product Brochure.docxTM65UXCL(the cross-connect, once).../Material/6fffa7f348f9475081b17eb1fc591b10/M1T1A669N837008945971838990/OptiXtrans E6616X Product Brochure.pdf(R025, 4 pp)So the brochure confirms the frame and the other direction of the trap: the R025 brochure prints OTN 680 Gbit/s, while the R026 docx prints 780 Gbit/s "only when a license is used" on
TM65UXCL, and the family page's E6616X column givesMaximum Number of Wavelengths: N/Awith100 Gbit/s (grey)per channel — a grey-light shelf with no DWDM layer, i.e. every port in it comes from a client board. Both brochures stop there.Also corrected on this thread: the E6616X is on the family spec table —
https://e.huawei.com/en/products/optical-transmission/optixtrans-e6600(200, 339,756 B) has four columns, E6616 / E6608 / E6608T / E6616X. Only the E9624X is the one absent from its family's table.Still the exact missing field: one client-board part number valid in the E6616X subrack plus its front-panel port census (count, rate, cage) — e.g.
TMK1EX10orTMB1LDCD, whichever the X accepts. Unblocked by: a readable copy ofEDOC1100156575BEJ09044(boards overview, 302 for every UA re-measured today), the E6616X Hardware Description (EDOC1100558547andEDOC1100625831both/view/notFound), or this issue naming the board to model (E6616X DC with 14x <board>).Re-assessed 2026-11-17, with the specific question this thread had left open: does the frame have a stated typical configuration the way the NetEngine 8000 M14 does? It does not — and this time the E6616X's own documents were counted rather than sampled. Issue stays open. Note: PR #1903 /
research/optixtrans-stated-config-absent.md; closes nothing.1. The "typical configuration" in the E6616X brochure is an MTBF footnote. OptiXtrans E6616X V100R026C00 产品彩页 (
https://e-file.huawei.com/marketingcloud/pep/asset/20000001/Material/41f5e9d76d5944f98b0f17008b050765/M3T1A590N1296407376208097515/OptiXtrans%20E6616X%20V100R026C00%20%E4%BA%A7%E5%93%81%E5%BD%A9%E9%A1%B5.docx, 200, 1,629,242 B): 典型配置 occurs once, in footnoteb— "以上参数根据产品典型配置计算得出" — attached to MTTR / MTBF 111.9 years / availability 99.999898%. Board part numbers in the text:TMB10,TMK10,TM651 (TM65UXCL, the system-control/cross-connect board). The R025 PDF brochure (.../6fffa7f348f9475081b17eb1fc591b10/M1T1A669N837008945971838990/OptiXtrans%20E6616X%20Product%20Brochure.pdf) has none. So the NetEngine route — a power row that lists cards with counts — has no E6616X counterpart.2. Every publicly readable document on the E6616X support page, counted.
https://support.huawei.com/enterprise/en/optical-transmission/optixtrans-e6616x-pid-257173887(200, 454,559 B) links 260 uniqueEDOCids. Requested one at a time at/enterprise/en/doc/<ID>: 47 answer 200 (23 of them product video), 213 answer302 → /view/notFound. A body scan of all 47 forTMB1*,TMK*,TM65*,TN*,TNG1*returns the empty set — including the largest open document,EDOC1100605035Huawei OptiX OSN 1800, OptiXtrans E6600, and DC908 User Guide (551,070 B), which is scope and procedure text.3. The chassis-side documents carry chassis codes only. OptiXtrans E6616 Datasheet (
.../621bfc22e8564cf69ae86581569a7594/M3T1A669N1304588489052958777/_2_-OptiXtrans%20E6616-Datasheet.pdf, 1,535,511 B):TNG10; its 26TMKcodes are the assembly chassis (TMK5K31AFB–TMK5K36AFB),TMK5FAN,TMB1APIUand the system-control/cross-connect set (TMK5SXCH,TMK5XCH,TMK6XCH,TMK5UXCME,TMK1XCL,TMK1FXCL). Section 1.4 explains slots in terms of "line boards or OTU boards" and "regeneration paired slots" — a board class, never a part number.4. The gated documents, re-measured today:
EDOC1100156575BEJ09044(E6600 boards overview) → 404;EDOC1100558547andEDOC1100625831→302 → /view/notFound; the same forEDOC1100163901(E9600 Board Datasheet) andEDOC1100621488. A correction worth recording: the e.huawei.com sitemaps do not index this chapter set at all (all 32 sitemap files, 111,538<loc>, 4,151 document pages; none of the known OptiXtrans material hashes appears), so index absence is not evidence of non-publication.Exact missing fields, unchanged: (1) one service-board part number valid in an E6616X subrack; (2) that board's front-panel census — port count, rate, cage. Grey light at 100 Gbit/s and
Maximum Number of Wavelengths: N/Amean every port this shelf will ever have comes from a client board, so there is no fallback.What would close it: one line of stated configuration (
E6616X DC (14x <board>)) from an E6600 ordering/configuration guide, or a readable copy of the E6600 boards overviewEDOC1100156575BEJ09044together with a statement that the board is valid in the X — the E6600 board chapters checked so far carry no Valid Slots table, unlike the E9600 side.Catalogue correction landed for the one member of this family that is modelled - PR #1942,
Refsonly, this issue stays open.devices/huawei/optixtrans-e6616.jsonsaid "14 service slots" with no power qualifier. That is exactly the trap this thread's 2026-11-17 pass documented and did not apply: chapter 1 (E6616 Chassis), Tables 1-2, in the document the entry already cites, gives DC (TMK5K31AFB) 14 service slots, AC 1500 W (TMK5K33AFB) 12, AC 2400 W (TMK5K35AFB) 10 - APIU boards take AC power slots 1 and 3, or 1-4. The same 14x B1LDCD build therefore read as fitting an AC shelf four boards (16 client ports) larger than it does.label,descriptionandnotesnow name the DC chassis and print the AC figures; the port list is untouched (56x 100GE client + 28x 200G line). Its E6608 and E6608T siblings already carried the qualifier.#803 itself is unchanged: one missing artifact. A service-board part number valid in the E6616X subrack with its front-panel port census (count, rate, cage). No new route was found today, and none of the enumerated ones was re-walked -
research/optixtrans-e6600-chassis-chapters.mdsection 7 andresearch/optixtrans-board-chapters-public.mdsection 5 are the record. Unblock: one TMB1*/TMK1* chapter whose Valid Slots table names OptiXtrans E6616X, or a stated configuration ("E6616X DC with 14x ").clear counters#1945Re-checked 2026-11-18. Still blocked, but the search space is now counted instead of sampled - and one document this thread has been citing turns out to be unpublished, not gated. PR #1973 carries the evidence (
research/optixtrans-e6616x-anonymous-doc-census.md).The census. The E6616X, E6616 and E6600-Series support pages link 299 unique
EDOCids. Requested one at a time with the Googlebot UA thatsupport.huawei.comrequires: 47 answer 200, 213 answer 302 ->/view/notFound. The 47 are 38(Video)items, 8 chapters of Routine Maintenance of OptiXtrans E6600 Hardware (EDOC1100475300-309), the joint OSN 1800 / E6600 / DC908 User Guide (EDOC1100605035) and the EUCC cybersecurity statement. The user guide's 74,406-character text layer contains zero occurrences ofTMB1,TMK1,TMK5,TNG1,Valid SlotsorPort Spec- it is the web-UI manual. No hardware description and no board chapter appears anywhere in the 299, so a board chapter is not merely unlinked from the product page; it is absent from the document list.A correction to this thread.
EDOC1100156575BEJ09044(the Boards Overview thatdevices/huawei/optixtrans-e6616.jsoncites) is not "302 to the login gate". Its chapter paths redirect into the viewer, and the viewer's own content endpoint,/hedex/api/pages/EDOC1100156575/BEJ09044/01, answers 404No doc data found.anonymously. That document is unpublished to anonymous readers, so no cookie, UA or idPath work will ever fetch it - the route is closed, not slow.Packages are closed at the root too. The "Downloadable Product Documentation Package" roots - and R026, the first generation that could carry X board chapters - are all 302: V100R026C00
EDOC1100625826, V100R025C10EDOC1100556468, V100R025C00EDOC1100526008, V100R024C10EDOC1100456079, V100R023C10EDOC1100366068. The viewer's PDF export,/hedex/api/misc/<id>/pdf-dl-url, answers 400 for each.Two more reads, both empty of boards. The
M1T1A669...E6616X Product Brochure PDF (4 pp, 216,145 B) is a distinct, older asset from the 26H2 tree and prints only the specification table; the V100R025C10 docx is the same document. A reseller price catalogue (tanatel-comm.com/product/e6616/) separates the X's hardware for the first time - chassis assembliesTM65K31AFB(DC),TM65K33AFB(AC 1500 W),TM65K34AFB,TM65K35AFB, versus the E6616'sTMK5K31AFB, plus boardsTMK1UNS4MPandTMK1XCL01(40G fgOTN lower-order cross-connect) - but it prints no Valid Slots column and the host refused every connection today, so its rows are a search-index lead, not evidence.Still missing, unchanged: a
TMB1*/TMK1*board chapter whose Valid Slots table names OptiXtrans E6616X, plus that board's front-panel census (count, rate, cage). What would close it now: a stated configuration of the shape already used for E6608/E6616/E9612/E9624 -E6616X DC chassis with 14x <board>- or any board chapter whose Valid Slots row reads E6616X. The frame arithmetic is already vendor-sourced and waiting; only the card is missing.Re-checked 2026-11-18 11:0x UTC. The field is not unpublished - it is behind a login, and this thread's own census could not tell the two apart. PR #1976 (
research/optixtrans-e6616x-docs-gated-not-unpublished.md) carries it. Issue stays open; no spec written.The instrument was the problem. The 2026-11-18 census read a
302 -> /view/notFoundon/enterprise/<lang>/doc/<ID>as "unpublished". A gated document redirects to the same place, so "213 of 299 missing" was really "213 not anonymously readable"./hedex/api/navi/<ID>/first-level-treeanswers in three states, and the controls separate them:EDOC9999999999,EDOC0000000000No doc data found.EDOC1100447312(public guide)EDOC1100456166NOT LOGIN!The E6616X's Chinese page is a second, product-scoped document list.
https://support.huawei.com/enterprise/zh/optical-transmission/optixtrans-e6616x-pid-257173887- 200, 437,872 B - links 176 unique EDOC ids: 90 answer 200, 86 answer 401, none 404. The gated 86 are the technical set:EDOC1100303423, R023C00EDOC1100328735, R023C10EDOC1100367203, R024C10EDOC1100456168, R025C00EDOC1100525984, R025C10EDOC1100558549(newest listed);EDOC1100625827and R025C10EDOC1100556465;OptiXtrans E6616X快速安装指南 02EDOC1100299473, and the three E6616X release documentsEDOC1100376749/50/51- none of which appear on the English page.So the statement in this thread that the board chapters are "unpublished to anonymous readers ... the route is closed, not slow" is wrong, and so is "no hardware description is in that list at all". Both were inferred from a redirect that a gated document also produces. The document that carries the board census exists; the crawler cannot present an account.
What would close it, unchanged in shape but now cheap: a free Huawei enterprise account reading the hardware description - https://support.huawei.com/enterprise/zh/doc/EDOC1100558549 (E6600 V100R025C10 硬件描述) or the English https://support.huawei.com/enterprise/en/doc/EDOC1100456166 (V100R024C10) - and naming the E6616X's valid service boards, i.e.
E6616X DC chassis with 14x <board>. The frame arithmetic is already vendor-sourced and waiting (DC 14; AC 12 at 1500 W, 10 at 2400 W).The registry row for #803 now says that, and no longer claims the field is withheld. #130 untouched this pass.
Catalogue correction on the one modelled family member, re-read from the chapter this thread already trusts - PR #1983,
Refsonly, issue stays open.The blocker is unchanged and exact: per E6616X-valid service board, (a) part number, (b) validity in the X subrack and which slots, (c) front-panel port census. Nothing new was found for the X; the census in the 10:38 UTC comment stands.
What was wrong was in the spec, not the thread:
devices/huawei/optixtrans-e6616.jsongave "2.8 Tbit/s OTN cross-connect for the 14-slot shelf" as if the shelf had one fabric. Chapter 1 (E6616 Chassis), Table 1 Current subrack capacity - same document - gives TMK5SXCH 2.8 Tbit/s (the only one with SDH: 10 Gbit/s per slot, 140 Gbit/s VC-4), TMK6XCH 2.8 Tbit/s OTN, TMK5XCH 2.4 Tbit/s. The note now names the board and states the 2.4 Tbit/s variant, and the description's unsupported "400G per slot capable" is gone (the chapter caps a slot at 200 Gbit/s ODUk and never prints "400G"). Table transcribed verbatim inresearch/optixtrans-e6600-chassis-chapters.mds.9.Re-fetched today: https://e-file.huawei.com/marketingcloud/pep/asset/20000001/Material/621bfc22e8564cf69ae86581569a7594/M3T1A669N1304588489052958777/2-OptiXtrans%20E6616-Datasheet.pdf - 200, 1,535,511 B, 20 pp.
Still blocked, but two claims in this record were wrong and one of them changes where the next pass should look. PR #1986 carries the evidence (
research/optixtrans-e6616x-brochure-editions-disagree.md); it does not close the issue.The correction
Yesterday's census note closed with: "Both print the same specification table - 14/12 service slots, 680G OTN / 680G OSU / 160G+20G SDH / 800G packet - and neither names a board." Both halves are wrong. Re-extracting the same two asset URLs (crawler UA, no new route):
M1T1A669N837008945971838990(216,145 B, 4 pp)M3T1A590N1227914753962237967, V100R025C10 (1,628,703 B)TM*board numbersTM65UXCLThe older PDF also contradicts itself inside four pages (highlights 620G/820G vs table 680G/800G). That matters here because this issue's body reads "OTN 620G / PKT 820G / SDH 160G" - 620G is a highlights figure that no Huawei table states, in any edition.
Where the next pass should look
TM65UXCLis the first board part number any Huawei-hosted E6616X document has yielded, and its footnote is the only board datum in the family: "The TM65UXCL board supports a maximum of 780 Gbit/s OTN cross-connect capacity only when a license is used."The X's boards are in the
TM65order-code family - not the E6616'sTMB1*/TMB5*/TMK5*. So the standing next step recorded in the registry ("aTMB1*/TMK1*chapter whose Valid Slots table names E6616X") is aimed at the sibling's namespace. Asking for aTM65*service board chapter is the sharper request.TM65UXCLitself cannot settle this: it is a cross-connect/system board with no client ports, so it has no front-panel census to give.Measured dead this pass
info.support.huawei.com- indexed with the E6616X page, but an alias:/enterprise/en/...,/enterprise/zh/...and/hedex/api/navi/EDOC1100558549/first-level-treeall answer 302 withLocation: https://support.huawei.com/.https://e.huawei.com/en/products/optical-transmission/optixtrans-e6600(45,722 B) and.../hu/...(53,363 B): the spec table those pages show in a browser is client-side rendered, and neither body carries a singleTM[56]order code.Unblocker unchanged, one word narrower: a
TM65*service board whose chapter states the frame it fits and gives its front panel (count, rate, cage), or a stated build of the shape the catalogue already uses for the siblings - "E6616X DC chassis with 14x<board>".Re-checked 2026-11-18 14:0x UTC. Two new artifacts, one new dead end. PR #1988 (
Refsonly); #803 stays open.1. The 26H2 sibling brochure, read end to end.
OptiXtrans E6616 V100R026C00 Product Brochure- https://e-file.huawei.com/marketingcloud/pep/asset/20000001/Material/dc67ba7b606b46df975107406dafb0a1/M3T1A590N1296416246227165363/OptiXtrans%20E6616%20V100R026C00%20Product%20Brochure.docx - 200, 515,695 B, 8,191 characters of text. It publishes a fourth switch option the cited chassis chapter does not list: TMK5UXCME, 920 Gbit/s OTN, 920 Gbit/s OSU, 920 Gbit/s packet, 160 Gbit/s VC-4 SDH - i.e. a ~0.9 Tbit/s shelf, not 2.4-2.8 T. It also prints 400 Gbit/s as "maximum single-channel rate" where the V100R025C10 chapter caps a slot at 200 Gbit/s ODUk. Both are now indevices/huawei/optixtrans-e6616.json, recorded side by side rather than merged.For the X it is silent by census: 0 occurrences of "E6616X", 0 of "TMB1"/"TMK1", 0 of "TM65"; the 10 "TMK" hits are all K5/K6 system-control boards in the ASON notes. No service-board table, in the newest brochure either.
2. The one public index of E6600 board chapters is an index of dead links. https://www.actfornet.com/solutions/solution/OptiXtrans_E66000_Boards_Overview (200, 226,637 B) names ~40 chapters with port counts (
TMK1EX1010x10GE,TMK1TDC2x100G,TMB1LDCD4x100G->2xOTUC2). Every link is the same document:/hedex/pages/EDOC1100156575BEJ09044/...-> 302,docid=EDOC1100156575, tocLibId=BEJ09044, tocLibVer=01/hedex/api/pages/EDOC1100156575/BEJ09044/01/resources/...-> 404, 18 B,No doc data found./hedex/api/navi/EDOC1100156575/first-level-tree-> 404 same stringThird document id to answer that way (after EDOC1100126589 and EDOC1100625542), so the hedex content API is confirmed dead for this class of document, not just those two. The index is useful as a name list, and one negative datum matters: no TM65 service board appears anywhere in it - the only TM65 order codes on any reachable artifact remain chassis assemblies (
TM65K31AFB...TM65K36AFB).Blocker unchanged and exact: a service board (TM65* or other) whose chapter states its part number, its validity in the E6616X subrack and slots, and its front-panel census. Frame arithmetic is vendor-sourced and waiting: DC 14 (slots 1-14), AC 12 at 1500 W, 10 at 2400 W. Full evidence:
research/optixtrans-e6616-brochure-26h2.md.Re-checked 2026-11-18 14:5x UTC. The boards are now named - by an installation, not by the vendor - and one standing inference in this thread is wrong. PR #1995 (
Refsonly) carries it; #803 stays open.The artifact. Ningbo Rail Transit Line 4 communications maintenance tender, appendix 4 (bill of maintained equipment): 135 pages, 736,057 B, 91,554 characters of text layer, downloaded and read end to end today. Fleet line:
延伸段传输系统设备(E6616X)2 套 车站通信设备室. Items 32-36:(The E6616 block 13 rows above it reads
TMK5K32AFB+UXCME+AUX, so the document is not reusing one subrack's list for both.)Correction. The 2026-11-18 12:59 pass closed with "the X's boards are in the
TM65order-code family ... asking for aTM65*service board chapter is the sharper request".TM65UXCLis a cross-connect/system-control board andTM65K31/K32AFBare assembly chassis codes -TM65*is the frame prefix. The service boards in the one record that names any areB2*:B2PD1(2M processing) +B2DMS(2M interface), the pair whose assembly the price list callsTMB2TPS- "32xE1/T1 tributary board (1DMS+2PD1)", and whose short form this same document prints asTPS ... ①1*DMS+2*PD1. So the X is a 2M/PDH aggregation node in the field - the use case this issue describes - and the sharper document request isTMB2*/TMK1*, notTM65*.Blocker, now exact. A spares list is not a build: every row is qty 1, so four cards cannot be expanded into fourteen slots without inventing ten more, and neither
B2DMS's own port count nor the slot cost of the DMS+2xPD1 assembly is stated anywhere here.What would close it: the per-slot complement of one real subrack (commissioning record or NE inventory export: "E6616X, TM65K32AFB, slots 1-14: n x "), or a Huawei board chapter whose Valid Slots table names E6616X. Full evidence and the dead ends (Tanatel still refuses TCP 443):
research/optixtrans-e6616x-deployment-board-inventory.md.Re-checked 2026-11-18 15:2x UTC. The last unexamined host is now measured:
support.huawei.cnis the same CDN, not a second index, so there is no anonymous copy hiding behind the mirror. PR #1998 (research/optixtrans-e6616x-instrument-portability.md). Issue stays open; no spec written.Measured today, both hosts, same pass, Googlebot UA (a browser UA gets Akamai
403):.com.cn257173887250732937EDOC1100558547E6600 V100R025C10 Hardware Description (EN)/view/notFound/view/notFoundEDOC1100558549硬件描述 R025C10 (ZH)EDOC1100163902E6600 Brochure and DatasheetEDOC1100621358Heat Dissipation and Routine MaintenanceEDOC1100171248Commissioning and Configuration (NCE) 11EDOC1100605035joint OSN 1800/E6600/DC908 User GuideAnd the instrument does not travel.
https://support.huawei.cn/hedex/api/navi/EDOC1100558547answers 400error!, not the 401NOT LOGIN!that separates gated from absent. So the three-state test the previous pass recommended must be run onsupport.huawei.com; a.cn302 proves nothing about publication.Two byproducts worth keeping. (1) The English HD edition list is complete and stops at R025C10 -
EDOC1100558547,EDOC1100525982(R025C00),EDOC1100456166(R024C10),EDOC1100367199(R023C10),EDOC1100328736(R023C00),EDOC1100303425(R022C10) - while the same page carries an R026C00 Feature Description (EDOC1100625831) and R026C00 release notes. So R026C00 has no English HD: an account holder who wants the edition the 26H2 findings come from should open the Chinese tree, where the R026C00 package isEDOC1100625827. (2) A cached search snippet renders this product page as "Sorry, no document is found for this product." The live page carries 260 ids - do not read that snippet as "Huawei publishes nothing for the X".Unblocker, unchanged and exact: one authenticated read of the E6600 hardware description -
EDOC1100558549(ZH, R025C10) orEDOC1100625827(ZH, R026C00 bundle) - to get the per-board Valid Slots table forB2PD1/B2DMS/TMB2TPSand theTMK1*Ethernet set. Failing that, the per-slot complement of one real subrack (E6616X, TM65K32AFB, slots 1-14: n x <board>). Routes now closed on top of the 2026-11-17/18 list: the.cnhost, and any assumption that an unauthenticated reader can reach these files at all.Re-checked 2026-11-18 15:5x UTC. New route, new corroboration, same blocker - and now the blocker has a number attached. PR #2000 (
research/optixtrans-board-tables-in-presales-decks.md) records it;Refsonly.Huawei's channel training decks contain the E6600 board tables. Two were read (as far as their hosts allow - both render slides as images, so the text below is the index text of those pages):
https://www.scribd.com/document/931093259/HCSA-Presales-Transmission-V1-0-Training-Material-TMB2TPS / TMB1PL1D / TMB1SP3DE / TMB1TPSEagainst2 Slot / 1 Slot / 1 Slot / 2 Slot, with2 x 16E1/T1 / 2 x 16E1/T1 / 2 x 21E1 / 2 x 16E1; the E6616X tributary slide prints2 Tributary board B2TPS/B2OTPS/B1TPSEand32 x E1/T1.https://www.slideshare.net/ssuserc99286/optixtrans-e6600-main-slidepdf-TMB1TPS(DMS+PD1) 32xE1/T1,DMS 32 x E1/T1 electrical signals ... Two PD1 boards and one DMS board work together to provide TPS protection.What that settles for this thread: the 32 E1 ports belong to the DMS board and
B2*is the E6616X's tributary family - the first source outside the Ningbo maintenance list to placeB2TPSunder an E6616X, so the inference in the 14:5x note is no longer only an inference.Why it does not close the issue: the same assembly (
1*DMS+2*PD1, three boards) is priced at "2 Slot", and2 x 16E1/T1is printed against a 1-slot board as well, so no per-board count can be quoted. A spec would be choosing between 2 slots/32 E1 and 3 slots/32 E1 on no evidence. No field orsourcesentry rests on a re-hosted deck.Still needed (Huawei-hosted, any one):
TMB2TPSslot cost in DC service slots;B2DMSalone front-panel count; or a stated typical configuration for the E6616X.Reopened (2026-11-18 18:30 UTC). This request was closed at 16:04:48Z, two seconds after PR #2000 merged. That PR is a research note about the E6600 presales decks; its own body says it does not answer this request and that
devices/huawei/optixtrans-e6616x.jsonis unchanged. The tracker takes a keyword-plus-number anywhere in a pull request body as an instruction, so a body that disclaims the answer still recorded one.Nothing changed underneath: no spec exists for the E6616X, and
research/open-device-requests.jsonstill declares the missing field to be "the service-board part number valid in an E6616X subrack, and that board's front-panel census". Perdocs/RESOLUTIONS.mdthe only close a device request gets is one where the catalogue answers it and the closing comment names the file, so this returns to the open queue.Blocker unchanged as of 2026-11-18 (no re-walk needed): the E6600 hardware description is published but login-gated - 86 of the 176 document ids linked from the Chinese product page answer 401
NOT LOGIN!, including every 硬件描述 edition and the R026C00 bundle - and the R026 brochure, the numbered e-file bundle and the Ningbo deployment tender between them name the parts without a per-slot complement. What would answer it: one authenticated session readingEDOC1100558549(OptiXtrans E6600 V100R025C10 硬件描述, newest edition), or a commissioning record / NE inventory export giving one real subrack's slot contents (E6616X, TM65K32AFB, slots 1-14: n x <board>).Re-checked 2026-11-18 19:4x UTC. The three chapters the last note named are mirror-only - and the criterion this thread has been carrying was aimed at a section the E6600 family never prints. PR #2022 carries the evidence (
research/optixtrans-e6616x-2m-pair-chapter-format.md);devices/is untouched.What was walked. The e-file route that produced
(6)-UXCMEand(21)-TMB3AUX, with the candidates the 14:5x note named. Five queries -site:e-file.huawei.com TMB2DMS,"TMB2DMS" site:huawei.com,"TMB2DMS" filetype:pdf,"TMB2DMS" "E6616X","TMB2DMS单板" 面板 32路E1/T1 有效槽位- return no Huawei-hosted asset for TMB2DMS, TMB2PD1 or TMB2TPS. Only verbatim reprints of the board chapters.What the reprints carry (a lead, not a source).
https://www.szhuaxunjia.com/product/2298.htmlprints接口规格: 支持32路75Ω/120Ω的E1接口或者100Ω的T1接口and table 12 配套关系 listing E6608 (TMK2UXCL/UXCLE), E6616 (TMK5SXCH/TMK5UXCME, TMK5XCS) and E6616X with TM65UXCL, from V100R022C10SPC010;.../2297.htmlcarries the same rows and states the PD1 has no panel ports, so the DMS presents all 32. That is the first text that answers "32 E1 per what" and "is a 2M board valid in an X" at once - and it is on a reseller host, so no field rests on it.The correction. "A board chapter whose Valid Slots table names E6616X" is asking for something that does not exist. The E6600 chapters carry no Valid Slots table at all (zero occurrences in
(6)-UXCMEand(21)-TMB3AUX); compatibility is a 配套设备 table naming the subrack and the cross-connect board. That is the family's chapter format, not the E9600's.research/open-device-requests.jsonnow names the right artifact.Still missing, exactly: the TMB2DMS chapter on a Huawei host, for (1) the panel census as vendor text and (2) a slot cost - the chapter states no slot count, and the mirrored dimensions (DMS 193.80 x 40.10 x 205.90 mm vs PD1 96.40 x 19.80 x 161.50 mm) do not reconcile to a fourteen-slot complement; plus a line board, which no source for the X has yet named.
What would answer it: an authenticated read of
EDOC1100558549(OptiXtrans E6600 V100R025C10 硬件描述 - any of the six editions on the Chinese product page) section TMB2DMS; a Huawei-hosted PDF of that chapter on the e-file asset route; or a commissioning record / NE inventory export giving one built subrack with its line board. The spec that follows isOptiXtrans E6616X (1x TMB2DMS + 1x TMB2PD1 + <line board>).Re-checked 2026-11-18 19:5x UTC. Still blocked on the service-board census; one route is now closed at part level instead of product level. PR #2025 carries it (
research/optixtrans-infofinder-api.md§8). This does not answer the request.Two measurements, both against
info.support.huawei.com:POST /info-finder/hardware/v4/queryModelswith{"domain":"0","language":"en","keyword":"<kw>","toolType":"hardwarecenter"}returns200 {"errCode":0,"data":{"total":0}}forE6616X, and forTMB2DMSandTMB2PD1- the exact two 2M boards the 19:4x pass named as candidates. They are absent from the tool's part index, so no keyword wording will surface them there. (Note for the next pass: anydomainother than the code"0"returns403 errCode 10101 "参数取值不合法", which reads like a refusal but is a type error.)GET /info-finder/search-center/en/enterprise/<type>/-pid-<pid>/hardwarecenteranswers 200 for datacom pids - AR628023772823, CE8875260431197, S8700-4255229035- and302 -> /enterprise/indexfor the control999999999and for every optical pid tried: E6616X257173887, E6616250732937, OSN 9800 U327905372, K36255972187, P3223221349, E6600 Series266120513. The tool is a datacom index; the E6600 family is absent, not thin.Exact missing fields: (1) the service-board set valid in an E6616X subrack with ports per board; (2) the board complement of one real, configured subrack.
What would unblock it: a stated configuration of the shape
E6616X DC chassis, slots 1-14: n x <board>, the readable OptiXtrans E6600 V100R025C10 硬件描述EDOC1100558549(any of the six gated editions), or a Huawei-hosted PDF of aTMB2PD1/TMB2DMS/TMB3EGS4/TMK1XCLchapter on the e-file asset route.Re-checked 2026-11-18 20:2x UTC. Still blocked on the service-board census. One route is now closed at the instrument, not at a query. PR #2027 carries it (
research/optixtrans-e6600-site-search-index.md);devices/is untouched.Huawei's own search cannot answer this, and not because of wording. The on-site search endpoint every
e.huawei.comsearch page declares -POST /external/api/services/pep/services/api/dynamic/apiserver/searchdirectnew/op,{language, site, keyword}- returns three tenants only: 官网搜索-产品与解决方案, 案例, 视频. QueriesTMB2DMS,TMB2PD1,E6616Xeach answer 200 withtotal: 0in all three. A hit here is impossible by construction: the published E9600 board chapters (TNG2DAP,TNG2SRAPXF) would also score zero. The CN branch the same script builds (e.huawei.cn/.../searchdirectnew/cn) is a 404 "no router rule" ALB page;e.huawei.com/enterprisesearchis a 504 (163 B).Locale is not a second door - measured, so it stops being a candidate:
support.huawei.cn/enterprise/en/doc/EDOC1100558549,support.huawei.com/enterprise/ru/doc/EDOC1100558549,/zh/doc/EDOC1100558549andinfo.support.huawei.cn/hedex/api/pages/EDOC1100558549all answer 302 (100 B / 355 B), whilesupport.huawei.cn/.../optixtrans-e6616x-pid-257173887renders at 200 (454,139 B). The V100R025C10 Hardware Description stays gated on all three hosts.And the CN product page is not a census either:
https://e.huawei.com/cn/products/optical-transmission/optixtrans-e6600, 200, 353,474 B, holds zeroTMB*/TMK*/TN*part codes; its specification table is client-rendered, so it yields prose and no board row.Exact missing fields (unchanged): (1) the service-board set valid in an E6616X subrack; (2) per-board front-panel census - port count, rate, cage; (3) slot cost per board (the
B2TPB-style assembly is DMS + 2x PD1, so "32xE1/T1" may be one slot or three).What would unblock it: a stated configuration (
E6616X DC chassis, slots 1-14: n x <board>), a readableEDOC1100558549(or the V100R024C10 / V100R025C00 Hardware Description editions), or a Huawei-hostedTMB2DMS/TMB2PD1/TMK1EX10chapter on the e-file asset route. Modelling the frame's empty slots as a repeatable module would be an engine feature, not a spec file.Re-checked 2026-11-18 21:2x UTC. Still blocked; one substitution is now measured shut rather than assumed shut. PR #2039 carries it (
research/optixtrans-osn1800v-hd-also-gated.md).devices/untouched.Why the OSN side looked like a way around the E6600 gate.
TMB2DMS/TMB2PD1are sold as OptiX OSN 1800 V Pro boards, and the only public index of this generation's board chapters links them under a1800v/hedex package path -.../resources/1800v/hd/pkg/tmk1ex10/tmk1ex10.html- not an E6600 one. If any OSN 1800 V hardware description were public, the same chapter (with the配套设备table that names E6616X) would be readable there.It is not. Seven editions listed on https://support.huawei.com/enterprise/en/optical-transmission/optix-osn-1800-v-pro-pid-250732494 (200, 476,757 B, 391 links), one
GET /hedex/api/navi/<id>/first-level-treeeach, Googlebot UA:EDOC1100558550,EDOC1100525975,EDOC1100456158,EDOC1100303420,EDOC1100249290,EDOC1100192279(hardware descriptions) andEDOC1100558552(product description) - all 401NOT LOGIN!. Controls the same minute:EDOC1100447312200,EDOC9999999999404, so the instrument still distinguishes public / gated / absent. Also 401: the EnglishOptiXtrans E6616X Quick Installation Guide 02EDOC1100299472, printed under Quick Start on the support page - one document with the ChineseEDOC1100299473already on record, so Quick Start is not a second door.Exact missing fields (unchanged): (1) the front-panel census of a board valid in the E6616X; (2) the per-slot complement of one built subrack.
What would unblock it: one free Huawei enterprise account reading OptiXtrans E6600 V100R025C10 硬件描述
EDOC1100558549(§ TMB2DMS), or a stated configuration of the shapeE6616X DC chassis, slots 1-14: n x TMB2DMS + n x TMB2PD1. Reseller reprints of the chapter do print "32 x 75-ohm/120-ohm E1 or 100-ohm T1", but a re-hosted page is not a source and no field will rest on one.Re-checked 2026-11-18 21:5x UTC. The blocker is unchanged; what changed is how it gets checked - this pass replaced the hand-run
curlritual with a script that answers the same question in one command. PR #2043 carries it (research/huawei-gated-doc-probe.md,scripts/huawei-doc-gate-probe.js,research/huawei-gated-docs.json);devices/is untouched and this issue stays open.The instrument, measured once and now reused.
GET /hedex/api/navi/<id>/first-level-tree(crawler UA) answers in three states -200public,401 NOT LOGIN!the document exists behind the account wall,404 No doc data found.no such document. The/enterprise/<lang>/doc/<id>page cannot tell the last two apart (both302 -> /view/notFound), which is the misreading two earlier passes on this thread had to correct.Baseline over 30 ids: 27
gated, 2public, 1absent.EDOC1100625826,EDOC1100556468,EDOC1100526008,EDOC1100456079,EDOC1100366068,EDOC1100331258,EDOC1100300582EDOC1100558547,EDOC1100525982,EDOC1100456166,EDOC1100303425EDOC1100625831,EDOC1100558553,EDOC1100558558,EDOC1100163902,EDOC1100621358EDOC1100299472,EDOC1100299473EDOC1100625921,EDOC1100625925EDOC1100558550,EDOC1100525975,EDOC1100456158,EDOC1100303420,EDOC1100249290,EDOC1100192279,EDOC1100558552EDOC1100605035(joint OSN 1800/E6600/DC908 web-UI guide),EDOC1100447312EDOC9999999999Newly listed, not newly readable: the 26H2 documentation set is now on the portal as
EDOC1100625826(OptiXtrans E6600 V100R026C00 Product Documentation 01), withEDOC1100625831Feature Description andEDOC1100625925Upgrade Guide - the set that goes with the R026 brochures read on 2026-11-17/18. All three answer 401, so the newest edition does not move the blocker; it does shorten the unblock step, because that package root is now the first thing to read if an account becomes available (it supersedes the R025C10 idEDOC1100558549thatdocs/DEVICE_REGISTRY.mdpreviously named).research/open-device-requests.jsonand the regenerated registry row now say so.Still missing, exactly: per E6616X-valid service board - (a) part number, (b) validity in the X subrack, (c) front-panel port census. What would unblock it: one authenticated read of
EDOC1100625826(its board chapters print配套设备, notValid Slots), a Huawei-hosted PDF of theTMB2DMSchapter on the e-file asset route, or a stated configuration (E6616X DC (12x <board>)).node scripts/huawei-doc-gate-probe.jsexits non-zero the moment any of these ids stops answering 401, so the next pass does not have to re-derive that either.Re-checked 2026-11-18 23:5x UTC. Still blocked on the service-board census, but the sibling substitution is now closed with the vendor's own numbers, and the fetch instrument is repaired. PR #2054 carries both (
research/optixtrans-e6616x-grey-ceiling-and-own-spec-sheet.md,scripts/huawei-material-fetch.js).devices/is untouched and this issue stays open.The E6616X has its own spec sheet, and it is grey-only.
OptiXtrans E6616X Product Brochure(https://e.huawei.com/marketingcloud/pep/asset/20000001/Material/6fffa7f348f9475081b17eb1fc591b10/M1T1A669N837008945971838990/OptiXtrans%20E6616X%20Product%20Brochure.pdf) — HTTP 200,application/pdf, 216,145 B, md56f217c3b057a42e86d927a3d18aca7e5, 4 pp, 6,037-character text layer. Product Specifications (E6616X): 221.5 x 220 x 442 mm, 10.6 kg DC / 12.3 kg AC, 14 service slots DC / 12 AC, OTN cross-connect 680 Gbit/s, OSU 680 Gbit/s, SDH 160/20 Gbit/s, packet 800 Gbit/s, maximum rate per channel 100 Gbit/s (Gray light), line rates 2.5/10/100 Gbit/s (grey).So the frame is fully sourced and the population is not transferable.
devices/huawei/optixtrans-e6616.jsonis modelled as 14x B1LDCD — 4x 100GE client into 2x 100G/200G CFP2 DWDM line. The E6616X sheet caps per channel at 100 Gbit/s grey with no wavelength plan, and the product page (https://e.huawei.com/en/products/optical-transmission/optixtrans-e6600, 339,756 B, md50d66ab2750b69fc2597ee095ede950c1) agrees: itsMaximum Number of Wavelengthsrow isN/Afor the E6616X and itsMaximum Rate per Channelrow is acolspan="3"400 Gbit/scell plus one100Gbit/s(grey)cell — a flattened dump of that table reads the wrong column, so check the colspans.E6616Xoccurs 0 times in the 22,964-character text of(2)-OptiXtrans E6616-Datasheet(1,535,511 B, 20 pp), so the datasheet does not claim those boards for both models either.The leaflet class is now measured shut, not sampled.
TMBoccurs 0 times in the E6616X brochure; 0 times inOptiXtrans E6616 V100R025C10 产品彩页(.docx, 516,108 B, 3,327 chars — the tenTMKmentions are K5SXCH/K5XCH/K6XCH/K5UXCME/K5XCS/K1XCL/K1FXCL, all cross-connect/control/clock); and 0 times in the Railway Optical Communication Network 26H1 brochure (8 pp, 6,117,094 B), whose device table groups "E6616/E6616X" in one row and lists board types (EMR8, UXCMS, UNQ2) only.Exact missing fields (now one item, not a census): a service-board part number valid in an E6616X, and that board's front-panel port census. Frame fields no longer need sourcing.
Instrument repair, because this pass hit it:
scripts/huawei-material-fetch.jsansweredpage 403(428-byte body) on every material page today — the host fingerprints the TLS client, not the headers (curl reads the same URL with the same UA andaccept: */*as 200 / 337,802 B in the same minute; addingaccept-language,connection: closeor a browser UA changes nothing). It now fetches through curl when curl is on PATH. Re-verified on the record's own hashes:(2)-OptiXtrans E6616-Datasheet1,535,511 B and(21)-TMB3AUX482,177 B, identical to the figure inresearch/optixtrans-e-file-numbered-bundle.md. Without this, a 403 on the only anonymous board-chapter route would have read as "the chapter is gone".What would unblock it: one authenticated read of
EDOC1100625826(26H2 package; board chapters print配套设备), the E6616X Quick Installation GuideEDOC1100299472, or a stated configuration of the shapeE6616X DC (12x <board>).node scripts/huawei-doc-gate-probe.jsstill reports 27 gated / 2 public / 1 absent on its 30-id baseline, measured tonight.Re-checked 2026-11-19 00:4x UTC. Still blocked - but the public bundle is now fully characterised, and the last two unopened members are measured negatives. PR #2057 carries it (
research/optixtrans-e6616x-public-corpus-inventory.md+ registry refresh). The request stays open; no spec was added.What was left unread, and is now read. The 27-part anonymous
(N)-bundle one-file.huawei.comhad two members no note records as opened:(17)-OptiX OSN E6616 MIBsapplication/pdf, 5,602,882 B, 273 pp, 495,778 extracted charsE6616X,OptiXtrans,TM65,TMB1,TMB2,Service Board; no enumerated board type (board typeappears once, as a trap payload).slot/portoccur only in theboard=%d,subcard=%d,port=%d,path=%dindex prose(3)-OptiXtrans FAT-TESTAccess detail worth keeping: both asset paths are printed in the
attachmentUrlinput of their material pages, ande-file.huawei.comserves them to no UA and Googlebot but 403 to desktop Chrome - a desktop-UA probe calls these files nonexistent.The MIB route is also shut, for the same reason as the rest. Section 6.2 Obtaining the MIB Files says log in at
support.huawei.com, then WDM/OTN > Software > version > download the package. The compiled MIB - the one artifact that would enumerateTMB2DMSas a board type - is a login-gated software download, not an anonymous file.Blocker, unchanged and now bounded on every anonymous surface: the service-board set valid in an E6616X subrack, and the front-panel census of
TMB2DMS/TMB2PD1(including whether1*DMS+2*PD1costs one slot or three). Nothing anonymously published states it - the public bundle is exhausted, the hardware description is 401NOT LOGIN!, and the deployment tender gives qty 1 per board.What would unblock it: one authenticated read of
EDOC1100625826(V100R026C00 documentation package) orEDOC1100558549(V100R025C10 硬件描述, 配套设备 table naming E6616X), or a commissioning/NE-inventory record of the shapeE6616X, TM65K32AFB, slots 1-14: n x <board>. Then the model follows directly: chassisTM65K32AFB, 14 service slots (DC) / 12 or 10 (AC), boards as stated.Re-checked 2026-11-19 01:2x-01:5x UTC. Still blocked - but the second support host is now measured instead of hoped for, and the zh series list is counted. PR #2065 (
research/optixtrans-e6616x-cn-mirror-same-gate.md+ registry refresh). No spec added.The .cn mirror is the same door.
support.huawei.cnserves the E6616X page byte-identically to.com(decoded md581580b68, 437,872 B; only gzip framing differs), and its hedex gate answers the same way:EDOC1100558549(V100R025C10 硬件描述) andEDOC1100625831(R026C00 Feature Description) are 401NOT LOGIN!againstuniportal.huawei.cn, while the known-publicEDOC1100605035is 200./enterprise/zh/doc/<id>still does the 302 ->/view/notFoundthat a gated document also produces.The zh twin of the E6600-Series list is new and empty. 75 ids, disjoint from the English 189; 40 open / 35 gated / 0 dead, and every open one sampled is a popular-science post (
EDOC1100496833传送小百科-第4期,EDOC1100535528第13期,EDOC1100593014第20期,EDOC1100550968【知光漫谈】). Union 264 ids: 63 open, 21 of them video/【Optical Basics】, the two substantive ones already known (UI guide, EUCC).Missing fields, unchanged: the service-board part number valid in an E6616X subrack and that board's front-panel census (count, rate, cage). What would unblock it: one authenticated read of
EDOC1100625826,EDOC1100558549orEDOC1100625831, or a stated configuration of the shapeE6616X DC (14x <board>). The frame arithmetic is vendor-sourced and waiting.Re-checked 2026-11-19 02:0x UTC. Still blocked - but the surface that replaced hedex is now measured, and it gates the same content a second way. PR #2067 carries it (
research/optixtrans-e6616x-support-gateway-not-a-doc-index.md+ registry refresh). No spec added.The product page is no longer server-rendered.
support.huawei.com/enterprise/en/optical-transmission/optixtrans-e6616x-pid-257173887returns 200 but 6,539 B of Vue shell, byte-identical with a/knowledge-basesuffix; the document list is fetched from/supportgateway/supproductservice/v1/enterprise/. Every endpoint the app names was called:multi-language/init?pid=257173887mid: SUPE_DOC,hedexPath: /hedex/hdx.do,isOpenJudgeSpider: falsemulti-language/navi/tree?vocationIdAbsPath=<x>257173887,1and9999999999999- 24 site categories; the parameter is unused, so this is not a product treeauto-complete/doc-with-lang?keyword=<kw>"data":[]forTMB2DMS,OptiXtrans,E6600,E1616,TMB3AUX,TNG2DAP- empty for published chapters too, so empty is not evidence of absenceaggregation/doc/second-item(the page's own call)WSF check error(support-productservice-010001), whilesecurity/csrf-tokenreturns"data": null- an anonymous session can never sign itSo the picture is now complete on both generations: hedex answers 401
NOT LOGIN!per document id, and the SPA gateway answers 400 to the one call that would enumerate categories. The only anonymous vendor route to an E6600 chapter remains thee-file.huawei.comasset path inresearch/optixtrans-e-file-numbered-bundle.md.One real gain, on a product page. The Greek edition (
https://e.huawei.com/gr/products/optical-transmission/optiXtrans-e6600, 200 / 54,484 B to a Googlebot UA, 403 / 433 B to a desktop UA) is a fourth vendor-hosted edition and printsNumber of Service Board Slots: DC chassis: 14 / AC chassis: 12/10andEquipment Capacity 2.8T/160G H/20G L(8.0 kg, 200G/λ, 80 λ), with no E6616X column. The AC 12/10 split that this thread sourced from the chassis chapter now has a product-page source too; the two editions differ (80 vs 96 λ, 200 vs 400 Gbit/s, 8.0 vs 10.6 kg) and neither has a12-only row.Missing fields, unchanged: the service-board part number valid in an E6616X subrack and that board's front-panel census (count, rate, cage). What would unblock it: one authenticated read of
EDOC1100625826,EDOC1100558549orEDOC1100625831, or a stated build of the shapeE6616X DC (14x <board>).Re-checked 2026-11-19 02:2x-0x UTC. Still open, but the blocker has moved out of Huawei's document tree and into this repository - and that is a different kind of blocker. PR #2072 carries it (
research/optixtrans-e6616x-port-type-gap.md+ registry refresh). No spec was added.The census is settled and reads 32 per pair. Three independent pages agree:
szhuaxunjia.com/product/2298.html(42,382 B - 32 x 75/120-ohm E1 or 100-ohm T1),szdingwei.net/Product/6542134456.html(34,588 B - titled32路E1/T1电接口板) and the Tanatel price row03034FAA TMB2TPS 32xE1/T1 tributary board(1*DMS+2*PD1). The配套关系table those reprints carry names the X directly -TMB2DMS | E6616X | TM65UXCL | V100R022C10SPC010- which is the answer this thread has been asking for since 2026-11-10. A re-hosted page is not a vendor source (docs/MAINTAINING.md), so none of it is quoted into a spec.What blocks the spec now is the engine.
engine/media.jshas no E1/T1 port type and no E1 cage: the families are ethernet / wireless / console / dsl / pon / coax / powerline / serial, andserialresolves to the V.35 cage (serial-v35, 1.544 Mbps, cablev35) whilecoaxis an F-connector. A 32-port G.703 board would draw as a WIC-2T or a cable-TV socket, whichscripts/audit-port-types.jspolices. So #803 is in the same class as #807/#808/#810 - rank 2 ofdocs/plans/stated-configurations.md, 'a port type the engine does not have' - and this stream carries no engine changes.Two smaller corrections.
https://e.huawei.com/en/products/optical-transmission/optixtrans-e6616xanswers 200 with<title>404</title>(336,300 B, identical to the non-existent-e6608), so no product page of its own exists; and Tanatel answers again (200, 1,020,978 B) with six X chassis SKUs - K31/K32 DC, K33/K34/K36 AC 1500 W, K35 DC - meaning every AC SKU of the X is 1500 W and the 12/10 AC split is an E6616 property, not the X's.What would close it: (a) a build -
E6616X DC (n x TMB2TPS), 32 E1/T1 per group, with the slot cost of one DMS against its two PD1s stated; and (b) an E1/T1 port type plus cage inengine/media.js. Neither is a Huawei publication gap. The vendor original of the 配套 table is still worth one authenticated read ofEDOC1100558549or the R026C00 packageEDOC1100625826to lift that row off a reprint.2026-11-20 04:3x UTC - reopened. The close at 2026-11-19T02:41Z was the forge's keyword parser again, and the catalogue still has nothing for this SKU.
Evidence the close was wrong, measured this pass:
grep -ril e6616x devices/returns nothing,research/open-device-requests.jsonstill lists issue 803, andnode scripts/check-closed-issues.jsreported it under closed against our own record (nothing in the catalogue). Issue 130 is the only other request in that state and it is still open, so 803 was the queue's one invisible entry.One new surface, with the control this thread has learned to demand.
https://e.huawei.com/cn/products/optical-transmission/optixtrans-e6616xanswers 200, 348,308 B, but so does the non-existent/optixtrans-e6699- byte-identical, md5c366504bee8a, zero occurrences ofE6616Xin either. So the Chinese edition is a shell served for any slug, and a 200 there is not evidence of a page in either direction. Recorded so it is not read as a discovery next pass.The two blockers are unchanged and neither is research: (1) a build - the vendor brochure publishes the frame (14 service slots DC, 780G OTN, 100G grey per channel) and no service-board table at all, and the only panel census (32x E1/T1 on the
B2DMS/TMB2TPSpair) exists solely on reseller reprints; (2) a port type -engine/media.jshas no E1/T1 cage, and this stream carries no engine changes.Both need a decision rather than another probe: a stated build of the shape
E6616X DC (n x TMB2TPS)from the requester or a maintainer, and approval of an E1/T1 media family. Nothing was modelled from the non-X E6616.2026-11-20 05:0x UTC — decision list corrected, nothing modelled. PR #2261 (docs only).
No probe this pass: the vendor tree is exhausted (the EDOC chapters answer 401 to an anonymous reader) and the only 2M census lives on reseller reprints, which
docs/MAINTAINING.mdwill not let a field rest on. What this pass fixed is the plan a maintainer reads.docs/plans/stated-configurations.mdcarried the E6616X at rank 3, "the vendor never published it", and its step 6 said an OTN port type was the only engine decision left. Both are wrong for this request now: the frame is vendor-sourced, and the ports are 2.048 Mbit/s G.703 (75 ohm BNC / 120 ohm twisted pair) or 1.544 Mbit/s T1, so approving the OTN media family that E9605/E9612/E9624 wait on leaves this request exactly as blocked as it is today. It is re-ranked 2b, with a new step 6b giving the concrete shape of an E1/T1 family (three media, two cages, a connector-first classifier ahead of theserialfallback atmedia.js:853,typeleft asserialso no view learns a new type).So the queue needs two yes/no answers, and neither is research:
E6616X DC (n x TMB2TPS), 32 E1/T1 per group. No source states how many groups a 14-slot X carries, or the DMS's slot cost against its two PD1s.Both blockers otherwise unchanged; no spec file was added and no engine change was made in this stream.
2026-11-20 07:5x UTC — three negatives and no spec; PR #2275 carries them (
Refs #803only, no closing keyword).node scripts/huawei-doc-gate-probe.js --group e6600-> 20/20 documents still401 NOT LOGIN, "nothing changed since the baseline".TMB1*/TMK1*toTM65*(the prefix of the X'sTM65UXCLcross-connect). Two queries overTM65*board documents return no board asset at all; the only TM65 string on any indexed Huawei host is still that licence footnote in the V100R025C10 brochure.UXCME:15,16=TMK5UXCME, E6608UXCLE:7,8=TMK2UXCLE, E6608TTMT1SCC- and no service-board part number: zero matches forB1L*,TMB*,TN13*,CFP2,QSFP28.Missing fields, unchanged: every port. Port count, rate and cage per service board, and the slots it is valid in.
What would unblock it: a
TM65*service board chapter whose Valid Slots table names E6616X and whose front panel gives the census; or one authenticated read ofEDOC1100625826(R026C00 package,配套设备table). Without either, a stated configuration of the shape the siblings use - "E6616X DC chassis with N x<board>" - plus the board's port count would let a spec land.One further gate, on this repo's side: an E1/T1 path needs a port type
engine/media.jsdoes not have today (nobnc,rj48org703cage;scripts/audit-port-types.jspolices it), which is an engine change and out of scope for this stream. Recorded indocs/DEVICE_REGISTRY.md.2026-11-20 19:3x UTC — still blocked, but two more routes are closed by measurement. PR #2304 (
Refs #803, no close)./cn/quirk.e.huawei.com/<locale>/products/optical-transmission/optixtrans-e6616xagainst a control slugoptixtrans-e6699: es/de/tr/fr return 200, byte-identical to the control (325,313 / 323,642 / 328,556 / 328,984 B), ru/ja return 301, 165 B, identical for both; zeroE6616Xin all twelve bodies. No locale has a real product page.support.huawei.com/.../optixtrans-e6616x-pid-257173887(2024-03-02, 356,312 B; 2024-12-13, 641,154 B) contain zeroTMB/TM65/TMK/B1L/EDOC/Slot/100G..../doc/EDOC1100456166,EDOC1100558549,EDOC1100625826were never archived — CDX[]for all three.Missing fields, unchanged: every port — the board set valid in an X subrack, and per-board count/rate/cage.
What would close it, unchanged and neither of them research: (a) a stated build,
E6616X DC (n x TMB2TPS)at 32x E1/T1 per assembly, which also needs the1*DMS+2*PD1slot cost settled (2 vs 3); and (b) an E1/T1 (G.703, BNC/120-ohm) media family inengine/media.js— without it a 2.048 Mbit/s port classifies asserialand draws as a V.35 socket.No spec file added; no engine change in this stream.
jaylouisw referenced this issue2026-11-20 21:25:46 +00:00
maxis indexed and dropped, probes are unchangeable, and an IPv6 path has no trace #23602026-11-21 05:4x UTC — this one leaves the queue. PR #2373 sets
Fixes #803.The blocker recorded on 2026-11-19 was two decisions: a build, and an E1/T1 port type the engine lacks. Only the second was ever real, and it only binds the 2M build.
The E6616X's service board is published with both halves in one chapter.
TMB3EGS4(OptiXtrans E6600 Hardware Description), read from the verbatim partner copy https://www.szhuaxunjia.com/product/2302.html (200, 67,351 B):TMB3EGS4 | E6616X | TM65UXCL | V100R022C10SPC010— the subrack-validity row this thread has been asking for since 2026-11-10, naming the X's cross-connect too.TX1/RX1~TX4/RX4, 连接器类型LC或RJ-45, 网管接口序号1~4; port modes GE electrical 10/100/1000, GE optical 1000M, FE optical 100M; one slot, 61 W typical / 66 W max.So the spec is
14x TMB3EGS4= 56x FE/GE (DC chassis, 14 service slots), the same stated-build convention asoptixtrans-e6616.jsonandoptixtrans-e9605.json.l2-forwardbecause the board bridges (EPLAN/EVPLAN, 16k MAC, 4095 VLANs, STP/RSTP).The 2M build (Ningbo:
B2PD1+B2DMS, 32x E1/T1 per DMS) stays unmodelled and is stated as such in the spec's notes and inresearch/optixtrans-e6616x-stated-build.md: the engine has no E1/T1 port type, and anrj45-declared G.703 tributary cables at the wrong rate. A pass that adds ane1cage can add that build as a second spec without reopening this one.Rejected candidates, recorded so nobody re-walks them:
TMB4EGS8(8 ports, but TX5-TX8 need a 端口扩展 RTU, so four by default) andTMB4EFS8(FE/100M only); ActForNet's E6600 Boards Overview carries no subrack column and never printsE6616X.