By: Koby Reshef

400G is delivering on its promise of higher capacity fiber optic transport to address the ever-increasing demands for speed and connectivity across metro, short- and long-haul network backbone infrastructure and between data centers. These new capacity levels allow carriers to maximize the value of existing infrastructure while reducing operational costs. Of course, like any new technology, 400G has its challenges.

Upgrades from 10G to 40G to 100G and 200G do not require any changes to the existing carrier infrastructure; they all have wavelength spectral bandwidths that fit into existing 50GHz/100GHz grid optical filters. The 400G capacity over a single wavelength has a much higher baud rate, making it too wide spectrally to pass through fixed 50GHz grid reconfigurable optical add-drop multiplexers (ROADM) and 50GHz channel-spaced filters.

Traditional 50GHz channels spacing mux and demux are standard for 40G, 100G, and 200G networks, but these are not compatible with 400G. In addition, the amplification requirements to meet the link budget of 400G networks are much higher, which account for the losses and gains to the receiver from the transmitter through the fiber.

The entire 400G optical network infrastructure “chain” needs to be fine-tuned to handle these various spectral issues. The mux and demux, ROADM, optical amplifiers, coherent optical transceivers, and digital signal processing (DSP) equipment need to be upgraded to support these new requirements.

Power and cost need to be considered within the equation as well. The 400G fiber optic solutions available today generally consume significantly more power than previous technologies and are much more expensive. Only a few vendors control the market right now, and adopting their proprietary technology means vendor lock-in for the long-term because their 400G solutions do not support the latest multisource agreements (MSA) either.

Two standard approaches to 400G

Two standards dominate the 400G market: the Optical Internetworking Forum (OIF) 400ZR, and the OpenROADM multisource agreement.

The 400ZR implementation agreement has been designed to reduce the challenges and costs associated with network flexibility and high-bit-rate data center interconnect (DCI) requirements. The 400ZR standard creates an economical, footprint-optimized, and simple method for transmitting 400Gb Ethernet over DCI links for short distances, generally up to 80 km. The specification for 400ZR uses higher-order modulation, such as 160QAM and DWDM. Operators can mix equipment from multiple vendors within the same network for the first time with the 400ZR requirements for interoperability at high speeds.

The form factor of the optical module used to deliver 400G has strict power consumption limitations, so the 400ZR DSP has limited functionality and supports only the Ethernet client with its use of the OIF Concentrated FEC (CFEC). This means it’s lacking encryption, tunable lasers, and OTN rates.

Another variant of 400ZR – ZR+ – has been pushed by the industry, offering the smallest form factor to push 400G speeds but without the reach limitations of 400ZR. The ZR+ module allows a span of up to 200 km using more powerful signal processing techniques because it requires 15W more than the standard 400ZR specifications. It’s the compromise between the telecom and data center market requirements.

Generally speaking, the 400ZR cannot be connected over carriers’ metro networks and ROADM infrastructures. Therefore, the OpenROADM multisource agreement by the Open ROADM standards committee has defined a stronger Open FEC (oFEC), which enables carriers to expand metro networks with 400G wavelengths or links.

Going modular

Standards-based 400G pluggable coherent modules help address some of the challenges across the data transport market. For example, the short-distance 400G pluggable transceivers that use PAM4 require multiple lasers, which cannot operate at long distances.