100G DWDM2 QSFP28 80 and 100G DWDM QSFP28 120 modules application guide

We offer both the DWDM-100G-Q28-120 and the DWDM2-100G-Q28-80, and we also frequently get a lot of questions regarding these modules, their differences, and their specifications. So we decided to compare both of these modules. In this article, you will find all of the right answers to your questions.

You’re here to find out which 100G DWDM2 QSFP28 80 or 100G DWDM QSFP28 120 module fits best for your exact needs. We can help you with that. The key differences in the modules themselves are distance and hardware. Both of the modules are PAM4, which is an evolution from the two-state (NRZ) modulation. The DWDM2-100G-Q28-80 has a dual CS connector, whereas the DWDM-100G-Q28-120 module has a double LC connector and is a single-lambda module.

DWDM2-100G-Q28-80

• •An unamplified module can typically reach only up to 10km, solely based on fiber loss and dispersion.
• •Just like with the DWDM 100G PAM4 Double LC transceiver model with an EDFA and dispersion compensation, in theory, it could be possible to exceed 80km, but it depends on the network conditions.
• •Not having dispersion compensation can degrade performance if you’re going beyond the 80km mark, requiring you to do additional network optimization.

DWDM-100G-Q28-120

• •With an unamplified module, you can only reach up to 5km without any external amplifiers.
• •With an amplified module, it is possible to push the modules up to 80km (With EDFA).
• •With dispersion compensation and additional amplification, it is possible to exceed 80km, but performance will be impacted depending on the network design and the fiber conditions.

Are DCMs necessary?

Before we go deeper, it is vital for us to know what a DCM is. DCM stands for dispersion compensation module. It is a passive dispersion management element that can be inserted into an optical fiber network to control the overall chromatic dispersion. A DCM can be used to compensate for dispersion across the entire C-band.

It depends on the fiber length and integrity requirements. If the modules, or for that matter any modules, use coherent modulation with built-in DSPs, it’s not generally necessary, considering that the built-in DSPs can compensate for the dispersion, making an external DCM unnecessary.

In case you don’t have coherent technology, a DCM may be required beyond the 40-80km range, because signals like DWDM experience chromatic dispersion, and non-coherent rely on the DCM for signal correction at longer distances.

But, in case you’re using modules that don’t have coherent technology and you’re not sure if the distance won't degrade the signal due to chromatic dispersion, I would suggest that you get a DCM for better optimization and assurance that your network will be at the utmost quality.

So, if your modules have built-in DSP (common for the modern 100g DWDM QSFP28 coherent optics), a external DCM device would not be needed.

In the off-chance your module doesn’t have coherent technology, using a DCM would be necessary for distances that exceed 40km to maintain the signal integrity.

Platform support

Well, since the modules are QSFP28, that pretty much means that they are compatible with networking equipment that has QSFP28 ports and supports DWDM optical networking. These modules can be used in switches, routers, and transport equipment from the major vendors, provided they do support 100G DWDM QSFP28 optics.

Compatible Switches & Routers include:

• •Cisco
• •Juniper
• •Arista
• •Huawei
• •Nokia
• •Ciena

A must for compatibility – Switch or router must have QSFP28 slots, some platforms may require FEC or tunable DWDM configurations, and lastly, if the modules are being used in the longer distance links, ~80km, EDFA could be a necessity.

Coding

First, we have to understand what this so-called hybrid solution is: Customized (hybrid) coding refers to the situations when vendors originally did not have a specific optical transceiver hardware type in the supported transceiver list. For the full article, click here.

So, since it is a special product, we would use a 100GBASE-ER4 code, which is a hybrid solution. But mostly the coding solely depends on what system you are using for your modules, referencing a different article written by us:

“We recently had successful 100G-QSFP28-80 (ZR4) tests in Nokia routers (7210 SAS-Sx, 7750 SR1, and 7250 IXR). Originally, these Nokia routers did not support a 100G ZR4 module, but we made a hybrid solution, where hardware remains 100G-QSFP28-80 (ZR4) and code is 100GBASE-ER4 (40km). With such a hybrid solution, we can have ZR4 80km link performance, but the optical transceiver is recognized as ER4 40km and is accepted by the networking gear.”

So overall, it depends on what systems the modules are being used in. By testing and configuring hybrid coding, it can help utilize modules in unsupported systems.

EDFA

What even is an EDFA? Well, EDFA is now most commonly used to compensate for the loss of an optical fiber in long-distance optical communication. It’s an optical amplifier used in the C-band and L-band, where the loss of telecom optical fibers becomes lowest in the entire optical communication bands. Another important characteristic is that EDFA can amplify multiple optical signals simultaneously, and thus can be easily combined with WDM technology.

An EDFA can amplify multiple channels, up to 96 wavelengths in the C-band, in fact. But you have to ensure enough output power, noise control, and gain flattering for the most optimal performance, but of course, there are limits.

All good things must come to an end, which is why EDFA’s have a maximum output power, typically between +20 to +25. If you have more than 100G channels, it means that you get less power per channel, which in turn means that you may get a reduced signal quality, and also, overloading the EDFA can lead to nonlinear effects. You also have to take into consideration distance and dispersion, a single EDFA usually might be enough for a distance of 80-120km. Going over 120km is where you may require multiple EDFAs.

Parallel Lambdas

Also at Edgeoptic, we have a DDMD-40 module that allows us to transmit 40 parallel wavelengths (lambdas) over a single fiber pair, one for TX, one for RX. However, there are factors that you need to consider when deploying 40 channels simultaneously.

For one, your DWDM MUX is designed for 40 specific IUT-T grid wavelengths in the C-band (1525-1565nm), meaning each 100G DWDM transceiver must be tuned to one of these wavelengths. Also, take into consideration your module and make sure that it supports tunable optics.

If you need or want 40 channels, think about the power that it will consume, having 40 channels = High aggregate power, which in turn means that you might have to deploy an EDFA for long distances. The MUX/DEMUX introduces insertion loss, which is around (4-6db per unit). This ensures that power levels stay balanced.

Also, running 40 active channels increases optical noise, and it is possible that a FEC may be required. If used in longer distances, OSNR degradation can limit the reach if not properly managed.

Over longer haul links that exceed the distance of 100km, chromatic dispersion can cause signal distortion, and may require DCMs, but if you’re using transceivers that have coherent technology, then the built-in dispersion compensation will be sufficient enough and an external DCM won't be required.

DWDM MUX/DEMUX filters should have high isolation (~30 dB) to prevent channel interference, but older or lower quality MUX filters can cause crosstalk between adjacent channels.

So which is better?

There is no clear winner, each has its perks and drawbacks. But if you had to choose, it would come down to what it is that you need for your systems and architecture.

The key differences are the optical budget. The –120 module is ~28-30dB, whereas the –80 module has a budget of ~22-24dB. For the –80 module, an SD-FEC is required only for longer reach, but for the 120 module, an SD-FEC is required. The same goes with Amplification, for the –80 module, it is optional if not using max reach, but for the –120 module, it is required if you’re using the full reach, so the EDFA can compensate for signal loss over long fiber spans. Taking into consideration the DCM and dispersion compensation, the -80 module may work without it, but at –120, chromatic dispersion becomes ever so increasingly more noticeable. If supported, you could also use coherent optics.

And the thing we are all thinking about is that the –120 are of course more expensive and consume more power, roughly ~4-5W higher, than the –80 modules. So to put it bluntly, choose the -80 module for metro networks with low optical loss, and the –120 modules for long-haul networks, ensuring that EDFAs and DCMs are in place.

Alternative Solutions for 100G long-distance WDM connections

Alternative and in some cases cost-effective solutions to 100G DWDM2 80 and 100G DWDM 120 modules are:

• •100G CWDM Solution – based on Single Lambda CWDM QSFP28 transceivers with reach up to 40km;
• •O-band passive DWDM 100G solution – allows to transmit up to 16x 100G data streams in 20/30km without the need for signal amplification;