Building a solid foundation for open optical networking

Disaggregation and open optical networking (including 400ZR) are the hottest optical topics of 2021, but these major trends did not emerge overnight. Open optical networking really began with the introduction of the open line system (OLS) several years ago — first in hyperscalers and then, more recently, in communications service providers (CSPs).

The OLS is a form of disaggregation in which the line system elements (including amplifiers, filters, and reconfigurable optical add-drop multiplexers [ROADMs]) are purchased separately from the terminal/transponder systems. The openness aspect of the OLS has a couple of dimensions. Allowing transponders to be chosen independent of line opens the door for multiple suppliers where, in the past, there was only one. A further degree of openness can be achieved with interoperability among transponder suppliers themselves, though this level of interoperability is not required for a line system to be considered "open."

The OLS architecture can also be referred to as "partial disaggregation" or "horizontal disaggregation."

Why partial disaggregation with OLS?

Heavy Reading survey research going back to 2017 has shown that OLS are important to CSPs. By disaggregating the terminals and the line, operators realized that they can efficiently achieve several major optical networking objectives:

• Reduce vendor lock-in: Operators are no longer obligated to buy transponders from the line supplier
• Lower capex: The introduction of competition drives down prices
• Encourage faster innovation: After installing a line system that will last for a decade or longer, operators are free to cycle through best-in-class transponders as they emerge

Today, the 400G coherent pluggables revolution is a perfect example of the benefits of adopting an open line. While OLS adoption began before coherent 400G pluggable optics were introduced, operators that adopted an open line are in an ideal position to capitalize on the trend of mixing coherent pluggable modules from the many module suppliers that are rushing to compete in this area. The Optical Internetworking Forum (OIF)-defined 400ZR specification may be less relevant here, as it is designed for short reaches below 120km (and often well below that max). However, telecom operators are particularly interested in emerging 400G pluggable optics that will travel hundreds of kilometers in distance and will require underlying line systems, including ROADM networks. Examples include 400G ZR+, defined as higher performance modules in 400ZR form factors, as well as the OpenROADM Multi-Source Agreement (MSA), which defines higher performance 400G pluggable optics in a CFP2 form factor.

Lastly, the new IP over DWDM (IPoDWDM) renaissance benefits from the foundation of an OLS. IPoDWDM means the introduction of pluggable optics directly into the switch/router, a move that reduces both capex and opex and simplifies operations by eliminating DWDM terminal hardware. Although it has existed for decades, the appeal of IPoDWDM took a giant leap forward with the introduction of coherent 400G optics in "client" form factors that do not reduce overall system capacity. Specifically, these 400G form factors are QSFP-DD and OSFP. Again, while an OLS may be less of a factor in very short-span 400ZR applications, it becomes essential in the 400G ZR+ modules that appeal most strongly to telecom operators.

Key considerations for open architectures

At the high level (described above) the trend toward open optical networking is clear. But in the complicated world of telecom, this does not mean that the road ahead is without its set of challenges. As operators plan their open optical networking roadmaps, there are some key considerations:

• Openness requires a level of standardization, whether its official or, more often the case today, de facto. Promising work in open optical networks is being defined in the OIF, Open Networking Foundation (ONF), Telecom Infra Project (TIP), OpenConfig, Metro Ethernet Forum (MEF) and OpenROADM MSA, among others. A reference architecture alone does not ensure product compliance with endorsed specifications. With disaggregated architectures, extensive certification testing will be key to promote interoperability and to speed the process from procurement to live network turn-up.
• Given the multi-vendor environment defined by disaggregation, extensive interoperability testing is crucial before networks are deployed in the field. Staged network testing can be used to validate both equipment interoperability and the functionality of the end-to-end automation. This approach saves money while speeding network turn-up and eliminates a lot of potential headaches along the way.
• A third challenge posed by disaggregation stems from the lack of multi-vendor network design and simulation tools on the market.To fill this gap, thorough lab and field testing are required to validate that the network will work as expected.
• Software-defined networking (SDN) control for an open optical network is a challenge, as an OLS does not come with this functionality out of the box. Some suppliers may leave the control functions to the customers to develop themselves (a model preferred by some hyperscalers). Other OLS ship with their gear-specific management systems but require a higher order controller for end-to-end management and automation across the disaggregated terminals and the line. CSPs may opt for an independent third-party software supplier for multi-domain control.
• 400G ZR+ has limited interoperability. While 400ZR is an OIF specification, 400G ZR+ is not part of the OIF and is not on the same path to interoperability — though the OpenZR+ MSA has some solid industry support. Heavy Reading research is not conclusive that operators will require interoperability in ZR+, but operators must be aware that mix-and-match pluggability will be limited for any pluggables that are not OIF 400ZR-compliant.

Despite challenges, the good news is that CSPs have a growing ecosystem of partners to work with as they move along their own paths to open optical networking.