What an OTN layer adds to a grey link
A grey module from port to port is a single span with one rate on it. An OTN layer sits under that port and wraps each client signal in a container with its own monitoring, its own error correction and its own protection switching, so the transport can carry several client types over one wavelength and be managed as a monitored path.
In practice OTN appears where a network has to carry 100G and 400G services beyond a single grey reach, where protection must switch without the routers seeing a change, and where service performance has to be monitored per path. This guide covers the optics and line cards around that layer. The transport platforms are described on the OTN and DWDM transport platform page, and the rate families are collected on optics selection guides.
Why 100G and 400G are the working rates
100G is the rate that made coherent detection ordinary. One coherent line interface carries a client at 100G over a metro or regional span with amplification, and the same line rate is the building block for aggregation. 400G is the rate that made coherent capacity cheap per bit, and it arrives in two forms: a pluggable that sits in a router cage, and a line card that sits in a transport shelf.
The two answer the same question differently. A pluggable avoids the shelf where the cage, the cooling and the host software support it. A card is the answer where client ports are grey or several clients are muxed into one wavelength.
Recommended OTN optics and line cards
| Module or card | Rate | Media / wavelength | Reach class | Where it usually lands |
|---|
| 100G coherent line module | 100G | SMF, C band tunable | Metro to long haul classes | Line side of a 100G transport link |
| 400G ZR and ZR+ pluggables | 400G | SMF, C band tunable | Metro to regional classes | Router ports carrying wavelengths directly |
| 800G coherent pluggables | 800G | SMF, C band tunable | Metro to regional classes | New high rate capacity on an existing line system |
| 100G transponder and muxponder cards | 100G | Client side per service | Per card and mode | Shelves mapping several clients into one wavelength |
| 400G transponder and muxponder cards | 100G to 400G | Client side per service | Per card and mode | High rate shelves with grey client ports |
| 800G DCO line cards | 800G | SMF, C band tunable | Metro to regional classes | Core sites carrying 800G wavelengths |
Reach is a property of the mode and the measured span budget rather than a figure on a data sheet. The pluggable route is compared on coherent pluggable optics and the card route on 400G transponder and muxponder cards.
Client side against line side
An OTN link has two very different optical jobs. The client side is short, grey and rate specific. The line side is a single amplified wavelength that carries the aggregate. Keeping them apart is what makes an upgrade an addition rather than a rebuild: the client optics stay as they are, and only the line side changes when capacity grows.
FEC, latency and protection
Forward error correction is what allows a coherent link to close over a span that the raw signal would not survive. A stronger code buys reach or rate at the cost of latency, and an OTN path is specified with a latency budget that includes the code, the client mapping and any protection switching.
Protection is usually planned as 1+1 on the line side or per card, which keeps the client up while the transport switches. Check that the protection path carries the same measured loss as the working path, because a protection path that does not close is not protection.
Monitoring, alarms and the span budget
The reason to place a service into an OTN container is the monitoring that comes with it. Per service counters, per span receive levels and amplifier gain settings belong in the commissioning record, and a span running at the edge of its budget shows up there long before it fails. Where an amplified line is being planned, the span list with measured losses, amplifier positions and span count decides whether the link closes.
Line systems and passive modules are described on the DWDM and CWDM passive module and DWDM mux and demux module pages.
Coding, availability and the spare rule
Client and pluggable parts are coded for the host named on the request, tested on that model, and they are not manufactured by that brand. Spares are counted per link family on the line side, because one tunable part covers the whole band, and per part group on the client side. One spare line card per card family, one spare line module per link family and one client spare per coded group scales from an access site to a core site without a plan change.
How to choose, in order
- List the client services with their rates, counts and protection requirements.
- Get the span list with measured losses, amplifier positions and span count.
- Decide per site between a pluggable in a router and a card in a shelf.
- Pick the span class and the code strength that closes the chain with margin, and check the latency budget.
- Choose the client optics per service and confirm the host release at each end.
- Set the spare rule per link family and per coded group, and record the span budget.
Frequently asked questions
Do I need an OTN shelf for a 400G link?
Not always. Where the router cage supports the pluggable and the cooling budget allows it, a 400G coherent pluggable carries the service directly. A shelf is the answer where the client side is grey or several clients share one wavelength.
How strong should the error correction be?
Strong enough for the link. A stronger code either extends the reach a rate closes over or allows a higher rate on the same span, and it costs latency. Ask for the mode list rather than for a single label.
Is the client side affected by a transport upgrade?
No. Client optics are a distance question and stay as they are. What changes is the line side.
How much protection is enough?
Match the scheme to the service. Line side 1+1 is the usual answer for protected capacity, and it only works while the protection path closes on its own budget.
What do you need to quote an OTN project?
The client service list with rates and counts, the span list with losses and amplifier positions, the line rate and protection scheme, the site type per location, the host release at each end, and the quantity per part with spares on a separate line.
Send the client service list, the span list with measured losses and the site type per location, and we come back with the line rate and mode that close the chain, the part per position and the spare lines set out separately. Related reading: DWDM metro transport, long haul coherent 400G and 800G, DCI OTN platform and all optics selection guides.
