In the field of optical transmission networks, the wavelength-independent, direction-independent, contention-free (CDC) ROADM architecture has attracted widespread attention. This new architecture is designed to provide greater flexibility and simplify operations. Although the advantages of this technology and architecture are attractive, there is still a lack of in-depth analysis of the core applications and advantages of CDC ROADM networks, and it is difficult to see an evaluation of the cost of different CD and CDC ROADM node architectures in the market.

Let’s take a look at it based on the above two points!

Core Applications of CDC ROADM

First, let’s take a look at the three core applications of the CDC ROADM architecture: bandwidth provisioning, bandwidth on demand, and protection switching/restoration at the optical layer.

Bandwidth provisioning

The bandwidth provisioning application allows operators to pre-deploy and connect “pooled” transponders and regenerators to CDC ROADM nodes at key network nodes. When additional bandwidth or services are needed, operators can quickly establish A to Z wavelength connections in the network, using these pre-set transponders. The deployment of new services can be shortened from months to minutes, greatly improving business speed and reducing the involvement of technicians.

But bandwidth provisioning also faces challenges, such as the cost of “pooling” transponders and the involvement of technicians in customer-side service activation. If growth patterns are unpredictable, unused transponders or regenerators will be scattered across the network, becoming stranded assets. Over time, these assets may be utilized by network growth. Although CDC ROADM enables pre-connection of the network side of the transponder, it still requires technicians to connect and activate the optical interfaces on the client side, which may limit the operational savings of pre-connecting the network interfaces.

Bandwidth on Demand

Bandwidth on demand is another common reason to implement a CDC ROADM architecture. Most telecom services are fixed between two or more customer locations and run 24/7. For many years, the telecom industry has been exploring the feasibility of providing transmission services on demand based on time periods. For example, an enterprise customer may want a 10G wavelength service between a primary data center and a secondary data center, but only needs the service between midnight and 3 a.m. to perform nightly backups and remote archiving.

On-demand wavelength services raise many questions, including whether there is a viable market for these types of services and whether they provide a positive return on investment for operators. In addition to the added OSS/billing complexity, it is unclear whether selling wavelength services in fractional time slots improves the operator’s financial performance or simply reduces revenue as enterprise users move to a less expensive on-demand usage model than purchasing operator bandwidth to accommodate peak demand periods.

Optical layer protection switching and re-optimization

The most commonly cited application of CDC ROADM networks is to implement optical layer protection switching and optical layer re-optimization. With optical layer protection switching, operators can choose among several protection methods, including 1:1 protection, <50 ms switching for critical services, 1:n shared optical protection, and dynamic mesh recovery. In particular, 1:n protection allows operators to protect their optical layer services without reserving 50% of their network capacity for protection as with 1+1 protection. Mesh recovery is the ability to dynamically calculate and signal new routes in the event of a failure. Mesh recovery is often used as a backup to 1:n optical protection in case of secondary failures.

The optical layer presents some unique challenges for optical layer protection switching. The optical layer consists of the physical fibers, lasers, and modulated signals that run on these fibers. Optical impairments such as optical loss, dispersion, polarization mode dispersion, and OSNR must be measured and managed on every optical span and route.

At the physical layer, these optical impairments vary by fiber type, modulation scheme, span length, and overall optical path distance (i.e., OSNR). ROADMs employ techniques to compensate for these optical impairments. When a failure occurs, optical path analysis must be performed on all potential protection paths. The Path Computation Engine (PCE) can calculate the feasibility of each optical protection path in real time, but a simpler approach is to rely on an optical path reachability table.

The optical path reachability table defines all reachable paths for each A-Z node combination in the network. The optical path reachability table can be calculated offline, using vendor DWDM modeling tools, and uploaded to the PCE periodically. When a network failure occurs, the PCE selects a protection path from the available routes defined in the optical path reachability table.

In addition to optical layer protection, many operators are also interested in an optical layer control plane for periodic re-optimization of the optical network, sometimes called “optical defragmentation.” Optical layer re-optimization enables operators to periodically clean up their network routes, recovering up to 20% additional capacity. Over time, business changes lead to fragments of bandwidth scattered across the network. By re-optimizing network connections every four months, the network can be “defragmented” and stranded capacity can be recovered.

Image Comparison of Classic, CDG and CDCG architectures

ROADM consists of optical amplifiers, optical switches, multiplexers/demultiplexers, transponders and muxponders, realizing a complete and flexible optical transmission node, as shown in the figure below. Mux/Demux provides the connection point between the composite WDM layer and individual channels or wavelengths, which are implemented with transponder and muxponder units.

In a classic ROADM, the Mux/demux is a passive device implemented with AWG technology, essentially a prism that separates each wavelength to its respective input and output ports. AWG technology is widely used, reliable, and cost-competitive, but each wavelength is fixed to a specific physical port. Moving a transponder to a different wavelength or to a different degree requires manual intervention by a technician to unplug the transponder from one port and reinsert it to a different Mux/Demux port.

CDC ROADM allows wavelength reallocation without manual intervention, but at the expense of higher node complexity and cost. The two main architectures under consideration are CD and CDC. Both architectures use a common optical core but differ in the implementation of the wavelength drop side. Both CD and CDC ROADM support flexible grid channel spacing and are sometimes labeled with a subscript “G” as CDG or CDCG ROADM, and are also represented by Flex Grid.

In the core backbone network, the CDC ROADM replaces the broadcast and selection architecture used on the “classic ROADM” with a routing and selection architecture based on 1×20 dual WSS modules, as shown in the figure below.

In addition, the fixed-port AWG mux/demux is replaced by a flexible WSS architecture, allowing any Transponder or Muxponder to be assigned to any wavelength and sent to any WDM degree.

Recently, the optical transport industry has shown more interest in the CDCG ROADM architecture due to its additional “no contention” feature. One limitation of the CD ROADM is wavelength contention. Wavelength contention conflicts occur when wavelengths of the same frequency (i.e., color) are terminated from different WDM directions (i.e., east, west, south, and north). Since the CD ROADM does not allow “contention” on the drop side, careful network planning is required to ensure that the wavelengths dropped at a given node are assigned unique frequencies. For many operators, avoiding the additional planning and operational issues required by the CD ROADM is a reason to migrate to the CDC ROADM.

The CDCG ROADM uses 8×16 MCS modules to implement wavelength drops, allowing any wavelength to be dropped from any direction to any client-side port (no contention). The high insertion loss of the 8×16 MCS requires additional amplifiers for each add and drop direction. The colorless capability is achieved by leveraging the inherent “tuning” characteristics of coherent receivers. Each coherent transponder “tunes” its optical receiver to the provided channel, essentially blocking or filtering out all other channels present on the optical port. A significant issue with the CDCG architecture is that a large number of amplifiers or amplifier arrays are required on the drop side as the number of channels and node size increase.

Image CDC ROADM cost comparison

To compare classic, CDG and CDCG ROADMs, we performed a cost analysis based on a two-direction (2D) ROADM (east-west) configuration, 88-channel drop capability, excluding transponders. Many operators refer to this as the “first cost”, which is the cost of the ROADM network before any services (i.e. transponders) are deployed.

The figure below lists the main optical components included in the Classic, CD, and CDC ROADM models.

The figure below provides a comparison of the corresponding standardized costs of classic, CDG, and CDCG ROADM nodes.

The 1×20 WSS dual routing and selection architecture, combined with the 8×16 MCS, provides the flexibility to support colorless, undirected, contention-free, and mesh-free WSS network architectures. However, the additional optical components and amplifiers result in higher initial node costs.

For the purpose of cost analysis, the Classic ROADM configuration and pricing were normalized to 1.0, providing a baseline for comparing CD and CDC architectures. The results of the cost analysis show that the difference between similarly configured Classic, CDG, and CDCG ROADM nodes is approximately 2.5x, with the increase primarily due to the use of 1×20 WSS modules in the core and 8×16 MCS in the Drop. At lower channel counts (<44), the CDCG ROADM is slightly cheaper than the CDG ROADM, but costs rise due to the increased number of amplifier arrays required as the number of Drop channels increases.

While the initial cost difference between ROADM nodes is significant, any price difference is miniscule compared to the true network costs of deploying 100G channels. As 100G transponders or muxponders are added to the network, the cost of these devices begins to dominate the overall network cost.

The relative cost difference between classic, CDG and CDCG ROADM nodes drops to 6-10%, which is a relatively small difference considering the more flexible and capable optical network supported by the CDCG ROADM architecture. Based on the above model, the figure below shows the cost comparison with 100G transponders.

Overall, CDCG ROADMs provide additional network flexibility by enabling wavelength reallocation and rerouting, optical layer 1:n protection switching, mesh recovery, and optical layer defragmentation, all without manual technician involvement.

However, this additional flexibility comes with increased node complexity and cost. The increased cost is primarily due to the additional WSS modules, amplifiers, couplers, splitters, and MCS modules required to implement a CDC ROADM. The cost difference between a CDC architecture with 100G channels compared to a “classic” ROADM node is only 6-10%.

Between CD and CDC ROADM architectures, the industry seems to favor CDC ROADMs due to the “no contention” capability.

In addition, for nodes with drop capacity less than 4T (i.e., about 40 100G channels), the cost of a CDC ROADM network is slightly lower. When we include the additional cost of the 100G unit in the analysis, the cost difference between the three network options is within 6-10%.