IP routing for the 400GE era
Entering the 400GE era
Relentless demand for more capacity at a lower cost per bit is forcing network providers to constantly rethink and reoptimize their network designs. Besides delivering more capacity for consumer internet and ultra-high-definition (UHD) video streaming services, they must provide high availability and low latency for the mission-critical and massive machine-type communication services that the cloud and 5G will enable.
Rapid advances in silicon are fueling a new generation of compact, pluggable coherent 400G optics that open exciting new avenues for optimizing IP-optical network designs. This is the second blog in a series about 400G as the universal currency for IP-optical convergence.
Evolution of coherent optics
Until recently, technology advances in coherent optics have focused on improving transmission performance with increasingly sophisticated digital signal processing (DSP) algorithms. Enormous progress has been made in this area, with probabilistic constellation shaping (PCS), introduced by Nokia in 2018, taking capacity close to Shannon’s limit to enable network operators extract maximum spectral efficiency from their networks. So where to next?
Improving optical transmission performance remains important for long-haul and subsea coherent transport applications, where fiber is expensive and scarce. But the main innovation focus has shifted to improving the power, space and cost efficiency of short- and intermediate-reach coherent optics. The introduction of 5G radio and evolution to distributed service architectures with compute, storage and peering resources located in the edge cloud are concentrating the bulk of capacity demands on fiber access, metro transport and regional data center interconnect (DCI).
Figure 1 shows the progress made in developing high-density optics, comparing progress in short-reach client optics for routers, and for pluggable coherent optics. Traditionally, there has been a sizable difference in the port densities of short-reach (gray) router optics and coherent line optics. The introduction of 400ZR and 400ZR+ pluggable transceivers closes the gap and removes the density penalty of using coherent optics in the same router ports designed for short-reach client optics.
Figure 1 Evolution of coherent optics
The compact QSFP56-DD form factor offers tremendous port density but its power dissipation is nominally limited to 14.5 watts, although transceiver designs are now pushing this up to 20 watts. This constrains the capabilities and reach of coherent pluggables compared to designs for larger form factors such as CFP2 and line cards in WDM systems with coherent optics.
Network operators need a range of 400G transceiver options to optimally address the different capacity, cost, topology and reach requirements in wide area networks. Table 1 lists the various options and their key characteristics.
Table 1 400G transceiver options
400ZR was one of the first efforts to standardize an interoperable 400G coherent interface specification. Developed by the Optical Internetworking Forum (OIF) and released in March 2020, 400ZR is profile-optimized for high-density access and point-to-point DCI applications. It can deliver 400 Gb/s up to 40km over a single dark fiber span without external amplification and support up to 64-channel WDM in the C-band and up to 120 km with external amplification. Although 400ZR can be supported in various pluggable form factors, QSFP-DD is the most prevalent implementation choice.
In contrast, 400ZR+ is a related, non-standard interworking specification that targets higher optical performance with flexible 100G–400G line rates and longer reaches by leveraging multiple modulation types (16QAM, 8QAM and QPSK) and high-gain forward error correction (open FEC). In 400G mode and depending on fiber quality and aging, 400ZR+ can reach up to 600 km with external amplification, and even further using subrates. It can also traverse a limited amount of reconfigurable add-drop multiplexer (ROADM) nodes with external amplification on the add/drop interfaces, albeit with reduced reach.
400G multihaul digital coherent optics (DCOs) are a new category of modular transceivers that package coherent optics in a larger form factor (CFP2) that can be equipped in routers and/or WDM transponder systems. Multihaul DCOs also support 100–400 Gb/s line rates using QPSK, 8QAM and 16QAM modulation. They leverage higher launch power to achieve longer optical reaches up to 750 km and can pass multiple ROADM hops.
400G Anywhere transceivers are conventional optical transponders that are performance-optimized to maximize wavelength capacity and reach. They take the form of integrated line cards that reside within WDM optical transport systems. State-of-the-art optical transponders can deliver wavelengths of 100–800 Gb/s over thousands of kilometers by applying sophisticated DSP techniques and high-gain forward error correction. Optical transponders are typically deployed in combination with ROADMs for regional and long-haul networks where fiber connectivity is scarce and costly. They are usually monetized as managed wavelengths for multiple services.
Transitioning to 400GE
It will take time to build an ecosystem for 400GE coherent pluggable optics, as with any new technology. But commercially available 400ZR, 400ZR+ and 400G Multihaul products will start shipping in mid-2021. Market uptake will be facilitated by the large installed base of QSFP-DD and CFP2-capable router ports that can be readily equipped to support 400GE, and by the many potential applications in DCI, metro access and aggregation rings, and metro/regional core networks.
Network operators need flexibility and choice in transceiver types to optimize cost and performance for a given (sub-)network or link because of dependencies relating to fiber availability, quality, reach, link topology and service requirements. IP-optical coordination is critical for seamlessly deploying, operating and assuring these options throughout the network.
Figure 2 depicts the transition to the 400GE era and the IP-optical interworking options that will enable this. The present mode of 100GE operation for most, if not all, operators is depicted on the left. It uses gray client optics in combination with optical transponders. The 100GE era started roughly ten years ago with the transition of IP backbone links to 100GE. It triggered a major upgrade cycle of core routing platforms. Today, 100GE is a ubiquitous interface in every part of the network, and 4x 100GE interface ports are a popular breakout option for QSFP-DD connectors.
Figure 2 Transition to the 400GE era
Plugging a coherent 400G transceiver into a router eliminates the need for an optical muxponder or transponder in the optical transport system. 400Gb/s is ample bandwidth to cost-efficiently fill a single wavelength: Most optical transport networks currently carry 100 Gb/s or 200 Gb/s waves. A router equipped with a pluggable transceiver effectively becomes an Ethernet muxponder and can address simple, single-span router interconnect scenarios over point-to-point fiber. An all-coherent metro/regional network must include a combination of density- and performance-optimized transceivers with different pluggable form factors to address all connectivity requirements. Optical transport systems with in-line amplifiers are deployed in addition to extend reach, efficiently groom wavelengths and enable 1+1 fiber protection.
Hybrid IP-optical solutions will efficiently meet variable capacity and reach objectives in mixed deployment scenarios that combine 100GE and 400GE interfaces or have link requirements that are beyond the reach of coherent pluggable transceivers. Operators that are evolving their 100GE networks to 400GE are likely to operate in this mode for the foreseeable future because it leverages their current optical network investments while offering incremental cost savings through the use of pluggable coherent 400GE transceivers. They may also still require present-mode solutions based on transponders on long fiber routes and for traversing larger numbers of ROADM hops in packet aggregation rings.
The 400GE era presents an opportunity for network operators to rethink and reoptimize IP-optical networks, and 400G coherent optical technology will play a key role in many future deployments. The choice of whether to evolve and optimize existing deployments or make a fresh start with next-generation solutions optimized for 400GE will largely depend on the age and longevity of each current network.
The Nokia IP-optical networking portfolio offers the scope, depth and range of 400G implementation options that operators need to make these decisions and succeed in the 400GE era. Nokia is a leader in 400G routing and optical technology and has achieved several industry firsts, including:
- Launching FP3, the first 400 Gb/s-capable routing silicon, in July 2011
- Demonstrating the first 400Gb/s IP routing interfaces in February 2015
- Shipping the industry’s first commercially available 400GE line cards in July 2018
- Supplying the first commercial deployment of 400GE router interfaces in March 2019
The Nokia WaveFabric Elements optical portfolio expands the 400G ecosystem with new pluggable coherent transceivers and high-performance coherent subsystems designed to meet surging demands from 5G and the cloud. Launched in May 2020, it leverages a new generation of PSE-V coherent technology and integrated silicon photonics to power and push new benchmarks for transmission performance, cost efficiency and interface density.
The next blog in this series will discuss Nokia’s pluggable coherent technology in more detail.
400G everywhere webpage
Technology: PSE Super Coherent Technology
eBook: Beyond the limit: Coherent solutions for the next decade
Technology: FP4 network processor
Solution: IP-optical coordination
White paper: The 400GE inflection point