Almost every optical part number in this catalogue is an order code: a string that identifies a line item and tells you nothing about the module. Ciena is the exception. XCVR-S40V55 can be read character by character — a 10 Gigabit SFP+, 40 kilometres, 1550 nm — and once you know the convention you can specify most of the range from the string alone. Then you meet 160-9603-900, which is from the same vendor, describes a 400 Gigabit module, and tells you nothing at all.
That is because Ciena has three naming generations in service at once, and they overlap on the shelf. This guide covers how each one reads, what makes transport optics different from data centre optics, the four things that reliably go wrong — including two of our own records that we want you to check with us before ordering — and the 69 Ciena compatible optics and transport platform modules Carritech holds, from 155 Mbit/s to 1.6 Terabit.
One number frames the range and it is the opposite of what people expect. Eighteen of the 69 parts are 400 Gigabit QSFP-DD, and there are 800G and 1.6T parts in here too. Every other vendor range on this site is shaped like an enterprise network, with the bulk of the parts at 1G and 10G and a thin tail at the top. This one is weighted towards the top end, because transport equipment is where capacity is aggregated rather than distributed: a metro or regional platform carries everybody else’s traffic, so it buys the fastest interface available and keeps it in service for a decade.
The three Ciena generations, and the other transport vendors
| Family | Example | Origin | How it reads |
|---|---|---|---|
| NT… | NTTA03AA | Nortel | Nortel-heritage codings. Opaque, and the oldest strings still in service. |
| 160-9nnn-900 | 160-9401-900 | Ciena, legacy | Numeric. The hundreds digit groups by family, but it is not a speed ladder. |
| XCVR-[letter][nn][code] | XCVR-S10V31 | Ciena, current | Fully decodable. Rate family, reach in kilometres, wavelength. Decoded below. |
| XCVR-[descriptive] | XCVR-400G-SR8 | Ciena, newest | Plain English: form factor, speed, port type. The newest convention and the most readable. |
| ZXS-QDnn | ZXS-QDS4ZZZZ-00 | Infinera | QD is QSFP-DD; S, L o D is the port type; the digit is the lane count. |
| ON… / OTR… | OTR400Q56DD_LR4L | Ribbon | OTR400Q56DD_LR4L spells out speed, cage and port type; ON330118 does not. |
| OD-PP… | OD-PPMR4F410ST0 | NEC | Opaque, but MR4 e SR4 fragments hint at the port type. |
| 1061… / 1184561… | 1061800930-01 | ADVA and Adtran | Ten-digit and seven-digit numeric order codes with variant suffixes. Nothing readable. |
The Nortel strings are there for a reason worth knowing: Ciena bought Nortel’s Metro Ethernet Networks business and closed the deal on 19 March 2010, which is how a product line that began at Nortel ended up with Ciena’s support organisation and Ciena’s later part numbers while keeping its original ones on the module labels. The same thing is happening again one layer out: Nokia completed its acquisition of Infinera on 28 February 2025, so the four ZXS- parts in the list below now sit under the same roof as the Nokia codings covered in our Nokia compatible optics guide. As with every consolidation, a merger does not make an optic interchangeable — the identification strings stay separate — but it does mean one supplier conversation can cover both, and it changes who you chase for an end-of-sale notice.
The XCVR coding, decoded
This is the most useful thing in this article. The decode below is read off our own mapping table rather than a published document, so treat the reach and wavelength fields as reliable and the family letter as a strong pattern rather than a specification. The structure is XCVR- then one family character, two reach digits e a three-character wavelength code.
| Field | Values observed | Worked example |
|---|---|---|
| Family character | 0 e A at 1 Gigabit, S at 10 Gigabit, V at 25 Gigabit, Q for the QSFP family | XCVR-S10V31 is an SFP+ |
| Reach digits | 00 short reach or multimode, then 10, 30, 40 e 80 in kilometres | XCVR-A40Y31 reaches 40 km |
| Wavelength code | G85 e Z85 are 850 nm, Y31 e V31 are 1310 nm, Y55 e V55 are 1550 nm, D43 e D45 are CWDM channels | XCVR-S80V55 is 1550 nm |
Read across the range and it holds consistently. XCVR-S00Z85, XCVR-S10V31, XCVR-S40V55 e XCVR-S80V55 are the 10 Gigabit reach ladder — SR, LR, ER, ZR — in one readable sequence. XCVR-V00Z85, XCVR-V10V31 e XCVR-V40V31 are the 25 Gigabit equivalent. XCVR-010Y31 e XCVR-A10Y31 are the same gigabit 1310 nm design under two family characters, which is the one part of the scheme we cannot account for from our own data.
Two warnings go with the decode, and both are the kind of thing that only shows up when you try to use a scheme rather than admire it.
- Il
Qfamily does not separate 100 Gigabit from 400 Gigabit. XCVR-Q10V31 is a 100G QSFP28 LR4 and XCVR-Q10D01 is a 400G QSFP-DD LR4. Same family character, same reach digits, and only the last three characters differ. On a platform with both cage types in the same shelf, that is not a theoretical hazard. Decode the family and the reach from the string, and get the speed and the cage confirmed. - Il
160-series is not a speed ladder. Il93nnblock is CFP2,94nnis QSFP28 at 100 Gigabit,95nnis QSFP+ at 40,96nnis QSFP-DD at 400 and97nnis Fibre Channel — so the numbers go up while the speed goes 100, then 40, then 400. The91nnblock is worse: 160-9103-900 e 160-9111-900 are 10 Gigabit SFP+ modules while 160-9113-900 e 160-9115-900 are 100 Gigabit CFPs, two characters apart. Never infer a part’s speed from its neighbours.
And one place where the decode and our own record disagree, offered plainly because we would rather you heard it here. V31 means 1310 nm everywhere else in the range — five parts confirm it — but XCVR-V40V31 is currently mapped to a Carritech 25 Gigabit design whose record says 1550 nm. One of the two is wrong and they cannot both stand. If you are ordering that part, tell us the wavelength the far end expects and we will confirm the module against it rather than against the part number.
400 Gigabit on multimode is three different port types, and the names are one character apart
This is the section to read if you take nothing else from this article, because the transport range is where all three live and because two of our own records currently describe them wrongly.
There are three ways to carry 400 Gigabit Ethernet over multimode fibre, and they came from two different IEEE projects:
| Port type | Standard | How it works | Fibre |
|---|---|---|---|
| 400GBASE-SR8 | IEEE Std 802.3cm-2020, approved 30 January 2020 | Eight parallel lanes, one wavelength each | Sixteen fibres, MPO-16 |
| 400GBASE-SR4.2 | IEEE Std 802.3cm-2020 | Four fibre pairs carrying two wavelengths each — short-wavelength division multiplexing | Eight fibres, MPO-12 |
| 400GBASE-SR4 | IEEE Std 802.3db-2022, approved 21 September 2022 | Four parallel lanes, one wavelength each, at 100 Gigabit per lane | Eight fibres, MPO-12 |
The SR4.2 and SR4 rows are the ones that matter. They use the same connector and the same fibre count and they are not interchangeable: SR4.2 puts two wavelengths down each fibre at 50 Gigabit per wavelength, SR4 puts one down each at 100 Gigabit. An SR4.2 module and an SR4 module will plug into the same trunk and will not link to each other. The giveaway is the fibre grade rather than the connector — SR4.2 was designed for OM5 and the wideband multimode it needs to separate two wavelengths — and the IEEE project that produced it, the 400 Gb/s over Multimode Fiber Task Force, is two years older than the one that produced SR4, the 100, 200 and 400 Gb/s Short Reach Fiber Task Force.
Now the part that is ours. XCVR-400G-SR4.2 is an SR4.2 part by its own part number, and our record for it currently describes 400GBASE-SR4 with a single wavelength per lane — which is the other port type. XCVR-400G-SR8 is correctly described as using an MPO-16, but our record says it carries the signal “across eight wavelengths” with internal multiplexing, where SR8 is eight separate fibres and no multiplexing at all. Both records also cite IEEE Std 802.3bs-2017, approved on 6 December 2017, which brought 200 and 400 Gigabit Ethernet in but defines none of these three multimode port types.
We are correcting all of it. Until we have, tell us the fibre count in the trunk and the fibre grade — sixteen fibres, eight on OM4, or eight on OM5 — and we will confirm the port type before anything ships. For the record, ZXS-QDS4ZZZZ-00 e ZXS-QDS8ZZZZ-00 are the genuine SR4 and SR8 designs in this range and are specified correctly. MPO polarity explained covers the connector side, single-mode vs multimode fibre covers how to tell OM4 from OM5, and choosing 400G optics works through the whole decision. On a transport platform the honest answer is usually that you should not be using multimode at all: the single-mode DR4, FR4 and LR4 parts in the same group use half the fibre and go a hundred times further.
Three more things that catch people out
1. The SR10 trap, and here it is invisible from the part number
100GBASE-SR10, from IEEE Std 802.3ba-2010, approved in June 2010, carries 100 Gigabit as ten 10G lanes over a 24-fibre MPO. 100GBASE-SR4, from IEEE Std 802.3bm-2015, approved on 16 February 2015, carries it as four 25G lanes over a 12-fibre MPO. Different lane counts, different lane rates, different connector. They will not link to each other and they will not share cabling.
NTTA03AA is an SR10 module, and it is currently mapped onto a Carritech SR4 optical profile. What makes this one worse than the dozen other instances in our catalogue is that nothing in the part number says SR10 — the Nortel string gives you no clue, so a reader cannot catch the error the way they could with a part called CFP-100G-SR10. If your far end or your trunk is SR10, say so when you order and we will confirm the lane count and the MPO size before anything ships. If you are not sure which you have, count the fibres in the trunk: twenty-four means SR10. CFP, CFP2 and CFP4 sets out the honest options for a CFP line card, one of which is adapting the cage to QSFP28 and leaving the ten-lane world behind.
2. Two Fibre Channel parts may be 32G modules matched to a 16G design
160-9700-900 e 160-9701-900 are the only Fibre Channel parts in the Ciena range, and our own records are candid about them: they note that the Ciena part is commonly listed as a 16GFC / 32GFC module and that the match to our 16 Gigabit design is a backward-compatible one rather than a like-for-like one.
That is honest, and it is also a thing you need to act on. A 16 Gigabit module in a 32 Gigabit fabric will bring the port up at 16, because Fibre Channel negotiates downwards — the Fibre Channel Industry Association roadmap states that each speed “maintains backward compatibility at least two previous generations” — so the link works and the capacity does not. Tell us the fabric speed and we will quote the right generation: we hold both, and the 32 Gigabit parts are a separate range. 16G and 32G Fibre Channel SFP+ e Fibre Channel transceivers work through the generation rules properly.
3. Eight Ciena-coded records are missing from the brand listing, and three exist twice
Worth saying because you will notice it browsing the site. The Ciena brand page shows 48 products. The catalogue actually holds 56 Ciena-coded records covering 53 distinct part numbers, and eight of them carry no brand term at all — including XCVR-100G-LR1, XCVR-100G-ER4L, XCVR-Q80V31, XCVR-CFP-100G-LR4 e XCVR-OSFP-2X400G-DR8. They are real, stocked, tested parts; they simply do not appear if you filter by vendor. Separately, XCVR-400G-SR4.2, XCVR-400G-SR8 e XCVR-800G-SR8 each exist as two records. If a Ciena part number is not where you expect it, search for the part number rather than browsing the brand, or send it to our compatibility checker, which searches the whole catalogue. We are tidying the taxonomy.
One more in the same spirit: XCVR-OSFP-2X400G-DR8 is recorded as an 800 Gigabit “twin-port” OSFP module and its description describes it in two incompatible ways in the same paragraph — as two 400 Gigabit DR8 connections, and as using a pair of two-fibre, four-channel MPO-12 connectors. DR8 is a single eight-lane 800 Gigabit interface; 2×DR4 is two independent four-lane 400 Gigabit ones, and which you want is a fabric design decision rather than a speed upgrade. Tell us whether the port is configured as one 800G interface or two 400G ones and we will confirm the part against that. Breakout explained covers the mechanics and QSFP-DD vs OSFP the form-factor choice underneath it.
What makes transport optics different
Four things are true of a transport platform and not of a switch, and together they explain why this range looks the way it does.
- The optic frequently faces a line system whose internals you are not told about. Where a data centre link has your equipment at both ends, a transport interface may hand off to a carrier or to a managed wavelength service. The governing document for that case is ITU-T G.698.2, current edition November 2018, which specifies the single-channel interface at each end of a line system rather than the system itself. Get the launch power and wavelength tolerance from the circuit order, not from the building plan.
- The payload is often not Ethernet. The optical transport network framing in ITU-T G.709, in-force edition June 2020, wraps client signals for transport, and an interface specified for OTU4 is carrying a different frame from the one a switch port expects even where the optics are identical. That changes nothing optically and everything about which support matrix the part number appears in.
- Coloured optics are the normal case, not the exception. The frequency grid in ITU-T G.694.1, current edition October 2020, and the CWDM wavelength grid in ITU-T G.694.2, in force since December 2003, are where a transport engineer starts. CWDM vs DWDM covers the choice and why tunable optics reduce complexity covers the spares argument, which is stronger here than anywhere.
- Service life is measured in decades. There are CFP and CFP2 parts in the list below on hardware still earning revenue, which is why your optics will go end-of-sale before your switch does is the single most relevant thing on this site to a transport operator.
One consequence worth stating plainly, because it is a buying trap. A 1550 nm part is not necessarily a coloured part. XCVR-S80V55 is a grey single-wavelength 80 kilometre module, and the Carritech design it maps to currently has a description that calls it DWDM. It is grey. If you are planning a multiplexed system you need a channel, not a wavelength band, and if you are planning a single grey span you do not want to be sold a coloured part. Say which and we will quote correctly. 10G DWDM and CWDM SFP+ sets out the difference and 400ZR vs OpenZR+ covers the coherent options where a grey part will not reach.
Which cage takes which speed
| Cage | Speeds in this range | Typical placement |
|---|---|---|
| SFP | 100M, 1G | Service interfaces and management on access and aggregation platforms |
| SFP+ | 1G, 10G, 16G FC | The long-serving workhorse. Client interfaces across the whole estate |
| SFP28 | 10G, 25G | Current client interfaces |
| QSFP+ | 40G, or 4 × 10G broken out | Older aggregation and inter-shelf links |
| QSFP28 | 100G, or 4 × 25G broken out | The dominant transport client interface for a decade |
| QSFP-DD | 400G, and QSFP28 optics in the same cage | Current line and client cards |
| OSFP | 800G, 1.6T | Newest high-density shelves |
| CFP, CFP2 | 100G | Line cards still in service. Large, power-hungry, and not going away |
Two ordering notes. A QSFP-DD cage accepts a QSFP28 module and the reverse is not true; the QSFP-DD MSA describes the form factor as adding “an additional row of contacts providing for an eight lane electrical interface” to the existing QSFP mechanical envelope, which is exactly why the backward compatibility runs one way. The MSA names its generations by per-lane rate — QSFP-DD at 50 Gb/s per lane, QSFP-DD800 at 100 and QSFP-DD1600 at 200 — which is a cleaner way to think about a migration than the aggregate figure. And an SFP28 cage will usually run at 10G, but which lower rates are offered is a platform and software question — SFP+ vs SFP28 covers it.
Five checks before you order
- Decode what you can and confirm the rest. The XCVR reach and wavelength fields are reliable; the speed and the cage are not, because the
Qfamily covers both 100G and 400G. - Never read a speed off a neighbouring 160- number. The series groups by family and the families are not in speed order.
- At 400 Gigabit on multimode, establish the port type from the cabling. Sixteen fibres is SR8; eight fibres is SR4 or SR4.2 and the fibre grade decides which.
- At 100 Gigabit on multimode, establish SR10 or SR4 before anything else. Count the fibres: twenty-four means SR10.
- Say whether the wavelength is grey or coloured, and quote the channel if it is coloured. On a line system the channel is the product.
The Ciena and transport platform optics Carritech stocks
Sixty-nine tested part numbers across Ciena, Infinera, ADVA, Adtran, NEC, Ribbon, Calix and Ligent codings, every one coded for the platform it is going into and measured against its specification before despatch. Listed by speed and form factor below.
155M and gigabit SFP — 12 parts
Filed at 155M, 4 parts. 100-01660, 1184561PG3, XCVR-000G85, XCVR-A00G85. A note in the interests of accuracy: these are 1000BASE-SX gigabit modules currently filed in our 155M category — a classification error on our side rather than a claim about the module. They are gigabit parts and they behave like gigabit parts. If you need a genuine Fast Ethernet optic for an old service interface, ask for 100FX explicitly. See 155M and 1.25G SFP modules and browse the 155M duplex SFP range.
1.25G duplex SFP, 6 parts. 1184561P1, XCVR-010Y31, XCVR-080Y55, XCVR-A10Y31, XCVR-A40Y31, XCVR-A80Y55 — the readable reach ladder in miniature: LX at 10 km, EX at 40 and ZX at 80, with the XCVR strings telling you which is which. Browse the 1.25G duplex SFP range and the 1.25G transceivers page.
1.25G CWDM SFP, 2 parts. XCVR-080D45, XCVR-A80D43 — two coloured gigabit channels at 80 km, and the only parts in the range whose wavelength code is a CWDM channel rather than a band. Quote the channel when you order. See 1.25G CWDM and DWDM SFP and browse the 1,25G CWDM SFP range.
Copper SFP, 2 parts. 1184561P4, XCVR-A00CRJ. A copper module contains no laser at all — it is a complete 1000BASE-T physical layer in a cage designed for an optical module, which is why it draws more power and runs hotter. On a transport shelf these are almost always a management or service-handoff interface rather than a traffic-bearing one. See copper SFP modules and browse copper SFP and SFP+.
10G SFP+, 25G SFP28 and 16G Fibre Channel — 12 parts
10G duplex SFP+, 6 parts. 160-9103-900, 160-9111-900, XCVR-S00Z85, XCVR-S10V31, XCVR-S40V55, XCVR-S80V55 — the full reach ladder from SR on multimode through LR at 10 km and ER at 40 to ZR at 80. The port types come from IEEE Std 802.3ae-2002, approved at the June 2002 Standards Board meeting, and there is no forward error correction at 10G over fibre, which means an error counter on a 10G client port means exactly what it appears to mean. See SFP-10G-SR: the full specification and browse the 10G duplex SFP+ range.
25G SFP28, 3 parts. XCVR-V00Z85, XCVR-V10V31, XCVR-V40V31 — SR, LR and the 40 km part whose wavelength we would like you to confirm with us. 25 Gigabit Ethernet comes from IEEE Std 802.3by-2016, approved on 30 June 2016, and FEC configuration is the usual reason a 25G link stays down — a host setting rather than a module property, which both ends have to agree on. See 25G SFP28 transceivers e forward error correction explained, and browse the 25G SFP28 range.
16G Fibre Channel SFP+, 3 parts. 1061701131-01, 160-9700-900, 160-9701-900. Fibre Channel standards come from the INCITS Fibre Channel Technical Committee — known as INCITS/T11 until January 2022, and still called T11 by most of the industry — which is “responsible for the development of the Fibre Channel (FC) standards”. The optics are named SW at 850 nm on multimode and LW at 1310 nm on single-mode rather than SR and LR. Read the warning above about 160-9700-900 e 160-9701-900 before ordering either. Browse the 16G Fibre Channel range and the Ricetrasmettitori 16G page.
40G QSFP+ and 100G QSFP28, CFP and CFP2 — 24 parts
40G QSFP+, 4 parts. 160-9501-900, 160-9503-900, 160-9504-900, 160-9506-900 — SR4 on an eight-fibre MPO trunk and three LR4 codings at 10 km. 40 Gigabit Ethernet arrived with IEEE Std 802.3ba-2010, approved in June 2010. Most of these ports break out to four 10G interfaces where the platform supports the split, which on a transport shelf is frequently how a handful of 10G services get delivered from one card. See 40G QSFP+ transceivers and browse the 40G QSFP+ range.
100G QSFP28, 12 parts — the second largest group. 160-9400-900, 160-9401-900, 160-9401-901, 160-9402-900, 160-9403-900, 160-9404-900, 160-9433-900, XCVR-100G-ER4L, XCVR-100G-LR1, XCVR-Q10V31, XCVR-Q30V31, XCVR-Q80V31. Six reach and port classes in one group: SR4 on multimode from IEEE Std 802.3bm-2015; CWDM4 at 2 km; LR4 at 10 km; the ER4 Lite design at 30 km without FEC or 40 with it; ER4 at 40 km; and ZR4 at 80. Two parts deserve singling out. 160-9433-900 e XCVR-100G-LR1 are 100GBASE-LR1 designs — a single wavelength carrying the whole 100 Gigabit, standardised in IEEE Std 802.3cu-2021, approved on 9 February 2021 — which means two fibres and one laser instead of four of each, and it is the quiet best answer to a 10 kilometre single-mode span if the platform supports it. At 100G and above forward error correction is part of the interface and permanently on, so a healthy port shows a steadily climbing corrected-codeword count rather than zero. The reach choices are set out in 100G QSFP28 transceivers e 100G ZR4 vs LR4 vs ER4. Browse the 100G QSFP28 range.
100G CFP, 4 parts; 100G CFP2, 4 parts. 160-9113-900, 160-9115-900, NTTA03AA, XCVR-CFP-100G-LR4 e 160-9300-900, 160-9301-900, 160-9310-900, 160-9320-900. This is the largest CFP and CFP2 presence of any vendor range on this site, and that is the single clearest signal of what transport hardware is: these cages were specified when 100 Gigabit was new, and the shelves holding them are still carrying traffic. LR4 at 10 km and ER4 at 40 on both form factors, plus the SR10 part discussed above. Read the SR10 warning before ordering NTTA03AA. CFP, CFP2 and CFP4 sets out the four honest options for an ageing line card, one of which is adapting the cage to QSFP28. Browse 100G CFP, CFP2 and CFP4, the 100G CFP e 100G CFP2 ranges, and the 100G transceivers page.
400G QSFP-DD — 18 parts, the largest group in the range
1061800930-01, 1061800930-11, 160-9600-900, 160-9603-900, 160-9604-900, OD-PPMR4F410ST0, OD-PPSR4ED10ST0, ON330118, OTR400Q56DD_LR4L, XCVR-400G-SR4.2, XCVR-400G-SR8, XCVR-Q10D01, ZXS-QDD4ZZZZ-00, ZXS-QDL4ZZZZ-00, ZXS-QDS4ZZZZ-00, ZXS-QDS8ZZZZ-00. Six port types across eight vendors’ codings: DR4 at 500 m, FR4 at 2 km, LR4 at 10 km and ER4 at 40 km on single-mode, and SR4 e SR8 at 100 m on multimode. The single-mode parts are the overwhelming majority and that is the right shape for transport. 400 Gigabit Ethernet was standardised in IEEE Std 802.3bs-2017, approved on 6 December 2017, with the faster electrical interfaces current parts rely on coming later from IEEE Std 802.3ck-2022, approved on 21 September 2022 — the same day as the short-reach fibre standard that gave us 400GBASE-SR4. Read the 400 Gigabit multimode section above before ordering either of the SR parts. Remember too that a QSFP-DD cage will take your existing QSFP28 optics, which is what makes a staged migration possible on a shelf you cannot take out of service. See choosing 400G optics, 400G optical transceivers, the real TCO of 400G e 400ZR vs OpenZR+ for the coherent options, and browse the 400G QSFP-DD range, the 200G QSFP-DD range for the intermediate step and the Ricetrasmettitori 400G e Ricetrasmettitori 200G pages.
800G OSFP and 1.6T OSFP — 3 parts
800G OSFP, 2 distinct part numbers. XCVR-800G-SR8 is an 800GBASE-SR8 design at 50 m on multimode through an MPO-16, and XCVR-OSFP-2X400G-DR8 is the twin-port single-mode part discussed above, at 500 m. 800 Gigabit Ethernet port types come from IEEE Std 802.3df-2024, approved on 16 February 2024 by the 400 Gb/s and 800 Gb/s Ethernet Task Force. Note that our record for the SR8 part describes it as a QSFP-DD module while filing it as OSFP; it is an OSFP part and the cages are not interchangeable, so confirm the cage with us. See Ricetrasmettitori ottici 800G, 800G OSFP optical transceivers e 800G and 1.6T migration, and browse the 800G OSFP range and the 800G transceivers page.
1.6T OSFP, 1 part. LMS3831S-PC+ — a DR8 design reaching 500 m over single-mode through dual MPO-12/APC connectors, and the fastest module in this range by a factor of two. Worth knowing where the standards are: 1.6 Terabit Ethernet is not yet a published IEEE standard. It is in progress in the IEEE P802.3dj task force, titled the 200, 400, 800 Gb/s and 1.6 Tb/s Ethernet Task Force, so today’s 1.6T parts are built to the OSFP1600 MSA and to pre-standard expectations. Buy them as a matched set with one supplier behind both ends. See 1.6T optical transceivers: what changes at 200G per lane and browse the 1.6T OSFP range and the 1.6T transceivers page.
One management detail that ties the top of this range to the bottom. Several of the newer records cite a CMIS-compatible two-wire interface. CMIS — the Common Management Interface Specification — is how a host talks to a modern pluggable, and it is now an Optical Internetworking Forum specification, adopted from the QSFP-DD MSA in January 2022 and covering QSFP-DD, OSFP, COBO and QSFP. The older parts in the same shelf use the SFF management layouts held in the SNIA SFF specifications library. Two management generations in one chassis, and a host that expects one will not read the other.
What the platform checks, and what it does not
What a transport shelf reads when a module seats is the identification data in the module’s EEPROM, laid out to the SFF or CMIS management specifications, and compares it against what its software release expects for that slot. If the strings line up, the interface comes up and the optic appears in the inventory like any other. If they do not, behaviour on transport hardware tends to the strict end: these are managed platforms where the module inventory is part of a configuration that gets audited and alarmed.
What the platform does not do is measure the optic. It cannot tell you whether the launch power is in specification, whether the receiver sensitivity matches the data sheet, or whether the part will still be in specification in three years — and on an 80 kilometre span with a thin budget, those are the only questions that matter. That is why every Carritech module is measured against its specification before despatch — see the 22 checks a module passes, which verify exactly the wavelength accuracy and receiver sensitivity a long coloured span lives or dies by. Spotting a failing fibre link before it drops covers reading the diagnostics once it is in, fibre optic connector cleaning is the cheapest decibel you will ever recover on a long route, and why optical link budget matters covers the arithmetic on ITU-T G.652 fibre, whose in-force edition dates from August 2024.
On the procurement questions, compatible optical transceivers vs OEM, OEM compatible vs OEM branded e do third-party optical transceivers void your warranty? answer them straight.
Inside the rack, do not buy optics at all
Transport shelves are frequently stacked with the switches and routers they feed, which makes this the easiest saving available. For a short link between a shelf and the router above it, a cable assembly is cheaper, draws no power, has no connectors to clean and no optical budget to manage. The direct attach copper e active optical cable ranges cover 10G to 800G, including 400G QSFP-DD DAC e DAC OSFP 800G — see DAC and AOC cables e DAC vs AOC.
One point that applies with particular force in a transport room: a cable assembly has an EEPROM at each end, and each end faces a different vendor’s expectations. A Ciena shelf at one end and a Cisco, Juniper or Nokia router at the other means the assembly needs correct identification twice, to two different firmware families. Plenty of suppliers can code one end. Coding both is what the Carritech Opticode coding box does at the point of deployment, and it is the difference between a cable that works and a cable that comes up on one side only.
Getting Ciena compatible optics without the OEM price
Carritech Optics holds the Ciena, Infinera, ADVA, Adtran, NEC, Ribbon and Calix coded range in UK and EU stock across the whole span: 155M e 1.25G through 10G, 25G, 40G e 100G to 400G, 800G e 1.6T, with the 16G Fibre Channel range, patch cords e media converters alongside.
Every module is coded for the target platform and measured against its specification before despatch, and carries a Garanzia a vita. The stock argument is at its strongest in a transport estate, and for a specific reason: this is the range where the hardware outlives the vendor relationships. A shelf bought from Nortel is supported by Ciena; an Infinera line system is now Nokia’s; and the CFP and CFP2 parts in the list above left volume manufacturing years ago. Holding one optical design per reach and coding at the point of deployment, rather than one line item per OEM string, is the answer to all three problems at once — the case is in building a smarter transceiver stock strategy, e your optics will go end-of-sale before your switch does is the piece to read before you plan a refresh. Our accreditations, why choose us e support pages cover how we work, and the knowledge centre holds the rest of the technical library, including data centre interconnect e how the optical transceiver industry is evolving. Our guides to Nokia, Arista e Brocade compatible optics cover the vendors most often found alongside.
Chasing a withdrawn Nortel coding, working out whether an XCVR string is a 100 Gigabit or a 400 Gigabit part, or trying to establish whether a 400G multimode trunk wants SR4, SR4.2 or SR8? Tell us the part number, the shelf or card model and the software release and we will say. Our compatibility checker returns the tested Carritech equivalent of an OEM part number in seconds and searches the whole catalogue rather than one brand listing, upload a whole part-number list and we will price it, or request a quote.