Cignal AI has updated its Transport Applications Report to include coverage of the ROADM line card market. Each quarter, Cignal AI surveys optical network equipment manufacturers who provide ROADM shipment data for the quarter. Three features will quantify the total ROADM market: port count (large, small), spectral coverage (C-band, L-band, etc…), and interconnection type (OXC or conventional).
The data is aggregated and presented in Cignal AI’s Transport Applications Report at the worldwide industry level. Individual vendor data and market share will not be included, but vendors can calculate their market share by comparing internal shipment data to the total market. Clients can download the latest Excel file that contains the new ROADM data from the Real-Time Excel download area of the report page.
Future analysis will examine the “fill ratio” – the comparison of aggregate shipped optical transponder bandwidth to shipped ROADM bandwidth of optical network sales. Such a metric can give insight into the state of infrastructure buildouts compared to capacity additions to existing systems and an indicator of the adoption of IP-over-DWDM in carrier networks.
The initial report includes data from all four quarters of 2021.
Coherent interfaces are the engine of optical network growth and the source of most optical hardware revenue. Continual advances in coherent technology have been responsible for the ability of optical networks to accommodate 30% traffic growth while keeping optical network capex in check with CAGR in the low single digits. With such criticality to the success of hardware vendors, network operators, and their networks, Cignal AI has tracked and analyzed coherent port shipments in all their flavors since its founding in 2016.
But while coherent interfaces are critical, they require a photonic layer infrastructure, or line system, to transport and route the wavelengths they generate from source to destination. This task falls primarily to Reconfigurable Add/Drop Multiplexers (ROADMs). Over roughly the same period in which coherent has taken over the task of modulation from direct detect transceivers, ROADMs assumed the job of photonic transport from fixed-wavelength filters. And in much the same way that coherent brought programmability to wavelengths, ROADMs have enabled programmability and its many benefits in the photonic layer.
- ROADMs provide topology flexibility – any-to-any wavelength connectivity independent of the fiber topology. They enable A-to-Z wavelength paths to match network traffic demands without costly regeneration.
- ROADMs provide resiliency. In combination with a distributed control plane or centralized controller, they can route wavelengths around network failures without the need for redundant stand-by transponders.
- ROADMs provide operational efficiency and reliability. With ROADMs, the optical layer is fully provisioned, requiring no additional touches or re-fibering to add capacity. New advancements such as MxN add/drop, line loading, and optical cross-connect chassis take this even further.
Line systems are increasing in importance given the growing prevalence of disaggregation in optical transport systems. Traditionally, optical transponders and line systems were sold as an integrated system, with long-lived line systems acting as a loss leader and vendor anchor, with margin recovery provided by ongoing sales of coherent interfaces as capacity is added to the network.
However, the transition towards disaggregated compact modular hardware has decoupled these two network functions. Combined with advances in open API’s and multi-vendor management, network operators are beginning to procure line systems and coherent interfaces independently. A vendor’s ROADM portfolio will increasingly be relied upon as a competitive differentiator and will need to stand on its own as a profitable product line.
IP-over-DWDM is another form of disaggregation which highlights the importance of standalone line systems and ROADMs. In IPoDWDM, optical transponders are replaced by router-based coherent pluggables, which relegates many optical transport systems to line system functionality.
Cignal AI will examine fill ratio trends in future reports – the ratio of aggregate shipped optical transponder bandwidth to shipped ROADM bandwidth. This ratio gives insight into the classic razor vs. blades comparison of line systems and visibility of longer-term capacity deployment trends. It should also allow IPoDWDM deployment trends to be calculated and quantified.
ROADMs are sometimes referred to as Wavelength Selective Switches, or WSS’s. But a WSS is more accurately considered a component of a ROADM, performing the core function of routing selected wavelengths to/from specific ports. It is surrounded by other components, including amplifiers and optical channel monitors. Some ROADMs contain two WSS’s, and a ROADM’s add/drop stage can also contain WSS’s. ROADM can also refer to an entire network element, or node, which performs reconfigurable add/drop functionality between multiple fiber degrees, or directions.
For this report, a “ROADM” refers to a single line card capable of terminating one “degree” or unique optical line interface. Low port-count ROADM line cards typically contain a single WSS component in the add path (an architecture known as broadcast and select). High port-count ROADM line cards nearly always contain two WSS components, one on each add and drop path (an architecture known as route and select). Both types are counted as one ROADM unit in this report.
Such cards are generally described as 1xN, with N being the total number of ports available for the interconnection of line-facing degrees or add/drop structures. (In reality, ROADMs are often 2xN, with the second trunk port used for path trace functionality). N also dictates the maximum number of degrees supported by a node built with N-port ROADMs.
ROADM line cards are always used in conjunction with other ROADM line cards and add/drop path components to form a complete ROADM node. For example, an M-degree node will contain M ROADM line cards. Some legacy ROADM architectures use a 1xN ROADM line card in the add/drop stage, but the quantity is negligible. A new generation of add/drop line cards, called MxN, utilizes the new MxN WSS component in the add/drop stage. Those line cards are not presently tracked as they are unique to add/drop and are not used as line-facing units.
Cignal AI’s ROADM reporting analyzes the total ROADM market along these three independent categories.
The port count of a ROADM line card determines the maximum number of degrees (optical lines) of a given ROADM node (the remaining ports are used for future degrees or add/drop components). ROADMs began as 2-degree devices and have continually advanced to today’s state-of-the-art 32 port line cards.
Few network architectures demand that many line-facing degrees, but high port-count ROADMs can facilitate the design of highly meshed networks, networks with multiple parallel optical lines, highly scalable add/drop capacities – or some combination of the three. With more ports comes increasing cost and complexity, and so lower port-count devices continue to play an essential role in networks requiring less connectivity.
ROADM line cards of 9 or fewer ports are categorized as “Small” ROADMs, while those with greater than 9 ports are categorized as “Large”. 9-port ROADMs dominate the Small category, although there is renewed interest in 2-4 port ROADMs for access and aggregation applications. 32-port ROADMs are prevalent in the Large category, supplanting 20-port devices, which not all equipment vendors chose to productize.
A DWDM system’s usable spectrum, the frequency range over which it can transport wavelengths, is determined by the bandwidth of its amplifiers and ROADMs. For most of the past 20 years, WDM networks have been deployed using the C-band, supporting 4.8THz of spectrum (or less) between 1529nm and 1567nm. Such systems support up to 96 50GHz channels (typically carrying 100G-200G wavelengths), 64 75GHz channels (typically carrying 300-600G wavelengths), or up to 40Tb/s using 800G wavelengths from Gen90P coherent interfaces over short distances. As coherent interfaces achieve spectral efficiencies approaching the Shannon Limit, interest has grown in expanding the spectrum over which WDM systems can operate to increase system capacity. There are two fundamental approaches – adding support for the L-band (1575nm to 1617nm) via a parallel set of ROADM and amplifier line cards, and expanding the existing C-band within a single set of ROADM and amplifier line cards. Combining both approaches is also feasible.
Three types of ROADMs characterized by their spectral coverage currently exist and are tracked by the report, with several others expected in the next several years
- 4.8THz C-band – the vast majority of ROADMs deployed. This category will also include legacy ROADMs supporting less than 4.8THz.
- 4.8THz L-band – an additional ROADM card added to C-band ROADMs in conjunction with band splitter/combiners and L-band amplifiers. In rare cases, L-band ROADMs are deployed independently of C-band ROADMs.
- 6THz C-band – these ROADMs use a single line card and associated amplifiers to support wavelengths from 1524nm to 1572nm, resulting in 25% more capacity than 4.8THz C-band ROADMs.
- 6THz L-band – Not yet shipping; these cards will be paired with 6THz C-band ROADMs, and support wavelengths from 1575 to 1626nm.
ROADM Internal Interconnect:
A third variable quantifies the percentage of ROADMs using an Optical Cross Connect (OXC). An OXC-enabled ROADM relies upon an OXC chassis instead of a “fiber shuffle” patch panel, and backplane connectors on the ROADM and add/drop line cards instead of faceplate LC or MPO connectors. Fiber patch cables among degrees and add/drop components are eliminated. OXC chassis feature an all-optical backplane and dust-proof, maintenance-free connectors and claim to significantly reduce footprint and cabling complexity compared to traditional ROADM implementations.
OXC ROADMs are currently supplied only by the three Chinese optical network equipment manufacturers. Most are deployed within China, although they are sold worldwide (except for North America). The vast majority of OXC ROADMs are of the Large category, but 9-degree ROADMs are beginning to be offered in OXC variants as the technology matures and finds wider adoption.
We’re excited about this addition to our Transport Applications Report. We believe it provides unique and valuable insight into the optical transport market. Like all our reports, it will evolve, and we would appreciate your feedback. Do these categorizations make sense? Are there other aspects of line systems you’d like to see reported? What additional analysis or data would be valuable? Please let us know at [email protected].