You are expanding your network. The goal is to ensure you won't regret this decision three years from now.

Start with four questions, not a product catalog

Before your first conversation with a supplier, answer four questions. They determine your choice of technology more than anything else.
01
How many kilometres are your longest links along the cable route?

The cable does not run straight. A route through a city, along buildings, with slack loops is often 20 to 40 percent more than the geographical distance.

Measure with OTDR, don't count from a map
02
How many channels do you need now and in five years?

CWDM has a ceiling of eighteen channels. DWDM starts with a few and scales to eighty or more without replacing line equipment.

Decides scalability
03
What throughput do you need per channel?

10G per channel is CWDM with classic modules. 100G without amplifiers on short routes is O-Band. 100G, 200G, and 400G on longer routes is coherent DWDM.

Determines the technology
04
Do you already have WDM infrastructure?

If you have a working CWDM, a CWDM and DWDM hybrid is often the fastest and cheapest path for expansion, without replacing cables or passive components.

May save you a replacement
The first question is: how many kilometers are your longest links along the cable route, not in a straight line? Cables don't run in straight lines. Routing through a city, through buildings, and including slack loops means the actual fiber length is often 20 to 40 percent greater than the geographical distance. Measure with an OTDR, don't count from a map.

The second question: how many channels do you need now and in five years? The architecture you choose today determines how easily you can scale in three years. CWDM has a ceiling of eighteen channels, while DWDM starts with a few and scales to eighty or more without replacing line equipment.

The third question: what are your bandwidth requirements per channel? 10G per channel means CWDM with standard modules. 100G without amplifiers over short distances means O-Band. 100G, 200G, and 400G over longer routes means DWDM with coherent modules. The fourth question: do you already have WDM infrastructure? If you have a working CWDM system, a hybrid of CWDM and DWDM is often the fastest and cheapest path for expansion, without replacing cables or existing passive components.

It is worth knowing the broader context of your options before focusing on choosing between WDM variants.

Port on an operator's switch

Simple and with no upfront investment, but every new link is a new fixed cost. The throughput is a single fixed speed, with no flexible scaling.

~3,500 PLNper month for a 10G port
Port on a muxponder (Alien OTN)

Leasing a port on an OTN device. About twice the price of a switch port, but with significantly greater aggregation of services and protocols.

5,000–6,000 PLNper month, OTU4 frame
Port on a multiplexer (Alien Lambda)

You lease a specific DWDM wavelength. Throughput is not limited by the port, from 10G to 400G on the same wavelength, with full protocol freedom.

10G–400Gon a single wavelength
Dark fibre

The greatest capability and full independence. With WDM a single fibre or a pair is enough, because you multiply throughput with technology, not with the number of fibres.

1 pairinstead of many fibres
Your own infrastructure with WDM

Full independence and maximum cost-efficiency at scale. The cost is the design and deployment, the gain is no fixed fees rising with every link.

full controlbest TCO at scale
The first is a port on an operator's switch. You lease a 10G port for about 3,500 PLN per month. It's simple, requires no investment, but every new link means a new port and a new fixed cost, and the bandwidth is a specific speed that you cannot scale flexibly. The second is a port on a muxponder, or Alien OTN. Instead of a switch, you lease a port on an OTN device that supports the transport protocol. Costs for an OTU4 frame range from 5,000 to 6,000 PLN per month—about twice as much as a single switch port—but with significantly greater capabilities for aggregating various services and protocols.
The third is a port on a multiplexer, or Alien Lambda. You lease a specific DWDM wavelength directly from an operator. Bandwidth is not limited by the port, so you can transmit from 10G up to 400G on the same wavelength, with full protocol freedom. The fourth is dark fiber or your own infrastructure, which offers the greatest capabilities and complete independence. Dark fiber lease rates depend on the route and the operator, but the key is that with WDM, you only need one fiber or a pair, because you multiply bandwidth using technology rather than the number of leased fibers. The fifth is your own infrastructure with WDM, providing full independence and maximum cost-efficiency at scale.

A checklist: ten questions before signing an order

Each of the following questions can change the specification or uncover a hidden cost.
01
Have you measured the real fibre length and attenuation on every route with an OTDR?

The straight-line distance is not the cable length, and each splice and connector adds insertion loss. Without an OTDR trace you are ordering against an unknown optical budget, and that is the single most common cause of a project that stalls after delivery.

Critical
02
How many channels will you need in five years, not just today?

CWDM tops out at eighteen channels, while DWDM scales to eighty or more on the same fibre. Sizing for today's traffic instead of the five-year horizon is what forces a full hardware swap far too early.

Key
03
Does the passive MUX and DEMUX pair have an expansion port?

A device without an expansion port is slightly cheaper, but expanding later means replacing the whole unit instead of adding modules. The price difference is a few hundred zloty, and the value at expansion time runs into five figures.

Critical
04
What signal power does the coherent module transmit at the multiplexer input?

Most DWDM systems are calibrated for 0 dBm at the multiplexer input. A module transmitting at around minus 10 dBm forces an extra EDFA amplifier per link, meaning added cost, higher power draw, and worse OSNR. Ask for the figure before ordering, not after.

Critical
05
Does your DWDM line system support flex-grid, and what channel width?

400G coherent modules need a channel at least 75 GHz wide. On a fixed 50 GHz grid the throughput drops to 200G. If you plan 400G on a single channel, confirm flex-grid support with your line-system vendor before you buy.

Verify with vendor
06
Have you calculated the OSNR link budget for each route?

Higher-order modulation demands higher OSNR, and the margin depends on distance and the number of amplifiers in the path. Assume nothing until the budget is calculated per link, so the modulation you order actually holds on your fibre.

Key
07
Do the router or switch firmware and ports actually support the module you want?

A QSFP-DD or OSFP slot fitting physically does not mean the host software will drive a coherent module. Check the vendor compatibility list and the required firmware version before ordering, because a firmware update in a production network is not a quick task.

Verify with vendor
08
Can one module type be programmed for different host vendors?

With the SRD environment you program a module for a specific vendor's hardware from a computer or smartphone, keeping one module type in stock instead of a separate version per vendor. Across a multi-vendor network this cuts both spare-parts cost and lead time.

Logistics
09
Have you built the module lead time into your schedule?

Standard modules are available off the shelf, but specialist configurations, LPO modules, and coherent 1.6T run to 8 to 20 weeks. With a hard launch deadline, order with a three-month buffer or the hardware becomes the critical path.

Scheduling
10
Is the system ready for future growth into the L-band or a hybrid?

Adding the L-band takes DWDM past 160 channels, and a CWDM plus DWDM hybrid multiplies capacity on existing infrastructure. Confirm the passive components and line system will not block that path before you commit to a specific build.

Future-proofing

Tools that simplify management: SRD 5 and GBC Photonics S100

Two items from our portfolio that simplify the lives of optical network engineers in practice.
Module programmer
SRD 5
Programming modules from a smartphone via the SRD Go app
Support for OSFP (400G, 800G, 1.6T) and QSFP-DD (400G, 800G)
Support for SFP-DD (100G) as well as SFP+, SFP28, QSFP+ and QSFP28
One module type in stock instead of a version per vendor
over USD 300,000in annual savings in a scenario calculated by Salumanus engineers, with equipment from three vendors across six network areas
DWDM platform
GBC Photonics S100
Signal power, wavelength, and OSNR available remotely via a browser or SNMP
The OCM card checks the status of DWDM channels without physical access to the device
The OLP module creates an automatic backup route and takes over traffic in the event of a failure
Modular 1U and 2U architecture, from 3 to 7 slots, buy only what you need
under 50 mstime for the backup route to take over traffic, with no restarts and no manual switching
SRD 5 is an optical module programmer that allows you to save thousands of zlotys on network management. You program modules from your smartphone via the SRD Go app, and the device supports OSFP (400G, 800G, and 1.6T), QSFP-DD (400G and 800G), SFP-DD (100G), as well as the commonly used SFP+ (10G), SFP28 (25G), QSFP+ (40G), and QSFP28 (100G) interfaces. Salumanus engineers have calculated that when working with equipment from three different vendors across six network areas, the SRD environment can generate over $300,000 in annual savings.
GBC Photonics S100 is a new platform for building DWDM systems that provides real insight into what is happening in the optical layer. Signal power, wavelength, and OSNR are at your fingertips, accessible remotely via a browser or SNMP system. The OCM (Optical Channel Monitor) card checks the status of DWDM channels without physical access to the device, and the OLP (Optical Line Protection) module creates an automatic backup route, taking over all traffic in less than 50 ms in the event of a main link failure, without restarts or manual switching. The modular 1U and 2U architecture with three to seven slots allows you to purchase only what you actually need.

Three mistakes we see most often

01
Buying without verifying the fibre

The module has a range of 40 km. The straight-line route is 35 km, but the real cable length is 47 km. Add the insertion loss of the MUX and DEMUX and the losses at splices. The system does not close the optical budget.

An OTDR measurement before the project takes a few hours. Debugging after deployment takes a week.
The most common mistake
02
No expansion port

A device without a port is a little cheaper. Expanding without a port in two years means replacing the whole device instead of adding modules.

The price difference is a few hundred zloty. The value of the decision is a five-figure saving.
Easy to avoid
03
Ignoring lead time in the schedule

Standard modules are available off the shelf, but specialist configurations, LPO modules, and coherent 1.6T mean 8 to 20 weeks of waiting.

With a hard launch deadline, order with a three-month buffer.
A scheduling mistake

How do we work with clients?

We don't start with a product catalog. We start by understanding your network: what you have, where it is, how much you are transmitting, and where it is going. Based on this, we calculate the optical budget for each route, recommend technology and components, and prepare a conceptual design with a quote and a five-year total cost of ownership calculator. All of this happens before any purchasing decision is made.

FAQ — Deciding on a Network Expansion

With four questions, not a product catalog. How many kilometres are your longest links along the cable route, not in a straight line? How many channels do you need now and in five years? What throughput do you need per channel? Do you already have WDM infrastructure? These four answers determine the choice of technology more than anything else. Without them, a conversation with a supplier comes down to browsing an offer rather than designing a network.
Because the cable does not run straight. A route through a city, along buildings, with slack loops means the real fibre length is often 20 to 40 percent greater than the geographical distance. That is enough to overturn the entire optical budget. A typical scenario looks like this: the module has a range of 40 km, the straight-line route is 35 km, but the real cable length is 47 km. After adding the insertion loss of the multiplexer and the losses at splices, the system does not close at all. An OTDR measurement before the project takes a few hours, while debugging after deployment takes a week.
Five, and each has a different cost profile. A port on an operator's switch is about 3,500 PLN per month for 10G, simple and with no investment, but every new link is a new fixed cost with no flexible scaling. A port on a muxponder, that is Alien OTN, is 5,000 to 6,000 PLN per month for an OTU4 frame, about twice the price, but with significantly greater aggregation of services and protocols. A port on a multiplexer, that is Alien Lambda, gives you a specific DWDM wavelength on which you transmit from 10G to 400G with full protocol freedom. Dark fibre gives you independence, and with WDM a single fibre or a pair is enough. Your own infrastructure with WDM gives you full control and the best cost at scale.
Rozpocznij z nami nowy projekt!
Dziękujemy! Otrzymaliśmy Twoje zgłoszenie!
Ups! Wystąpił błąd podczas wysyłania formularza.

Is there a universal optical module?

Have you ever chosen to run some kind of fiber optic cable and find yourself in a situation where, when you arrive at the installation site...
Czytaj więcej

400ZR vs 800ZR vs OpenZR+: Full Technical Analysis for DCI Network Engineers 2026

400ZR is the most widely deployed coherent standard in history. 800ZR is growing at 145% CAGR to 2029. Complete technical analysis — DSP fundamentals, operating modes, OSNR, economics per Gbps, and roadmap to 1600ZR.
Czytaj więcej

GBC Photonics S100 - nowa platforma do budowy systemów DWDM

GBC Photonics S100 to aktywna platforma DWDM. Umożliwia pełny monitoring warstwy optycznej, automatyczne przełączanie tras, intuicyjne zarządzanie przez WEB GUI i SNMP oraz elastyczną rozbudowę bez wymiany infrastruktury.
Czytaj więcej