R&D labs around the world are developing numerous networking technologies that will receive significant time, money, personnel, and attention throughout 2020. These technologies are aimed at allowing submarine network operators to better modernize their undersea network assets to maintain pace with voracious growth in bandwidth demand, provide competitive network service differentiation, and stand out in a hypercompetitive submarine networking market.
Spatial Division Multiplexing (SDM) cables
With voracious and ongoing bandwidth growth experienced for many years now, coupled with expanding rollout of 5G services that significantly increase access speeds to content hosted in data centers, cable operators are seeking new ways to increase available bandwidth between continental landmasses for Data Center Interconnection (DCI) purposes; satellite networks need not apply! Only fiber-optic-based networking technology can deliver the massive amounts of capacity and economies of scale required.
Although Submarine Line Terminating Equipment (SLTE) has been constantly innovative at a frenetic pace for several years now, the wet plants to which they connect have experienced comparatively little innovation—until now. Wet plants leveraging SDM technology offer more fiber pairs than traditional submarine cables. And although SDM cables support less capacity per fiber pair, they offer a much higher overall capacity due to far more pairs (12 fiber pairs and higher), which is further enhanced via power-optimized repeater (misnomer for subsea optical amplifier) designs.
As an industry proof point, the first SDM-based submarine cable deployed is the transatlantic Dunant cable, which supports up to 250 Tb/s of overall capacity over an aggregate of 12 fiber pairs—many more than the traditional six to eight fiber pairs offered on recent submarine cable deployments. SDM will continue to be the submarine cable technology of choice moving forward, performance and economics permitting.
Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL)
AI, ML, and DL are interrelated, and are making their way into all parts of the global network infrastructure—and submarine networks will not be the exception. These software-based technologies were born in the same data centers interconnected by submarine networks, which now carry over 50 percent of global submarine traffic between data centers dispersed around the globe.
So why not use these technologies on the same DCI networks that help enable them in the first place? Some network operators have already answered this question by deploying these technologies to optimize and monetize their expansive interconnected network assets, overland and undersea, for competitive advantage. This trend will continue since associated capabilities, such as preventive maintenance, are too hard to ignore.
“Garbage in, garbage out”: the old adage still applies to the above-mentioned AI/ML/DL technologies, so the network performance data used must be accurate, timely, and ongoing. This is achieved via real-time streaming telemetry from network equipment accessed via open Application Programming Interfaces (APIs). Who better to report on how the network is ‘feeling’ on an ongoing basis than the network itself?
Although this does make intuitive sense, operators will first need software interfaces to access the real-time performance data used by AI/ML/DL for actionable insights. However, access to open APIs is not enough. The network, on the other side of the API, must be highly instrumented to provide relevant, accurate, and timely performance metrics on an ongoing basis that’s accessible via these open APIs.
Highly instrumented networks, real-time performance data, and open APIs allow network operators to ensure their networks are working in an optimal manner to guarantee Service Level Agreements (SLAs) on an ongoing basis.
Open Submarine Cables
Everyone wants greater choice, but with it comes greater responsibility. Building a best-in-breed submarine network means selecting interconnected technologies from different niche vendors who are experts in their fields, since no single vendor is the best at everything required for a global end-to-end network. For example, an operator choosing the best wet plants and coherent modems for their Open Cable design and associated business requirements typically would work with both wet plant and SLTE vendors to ensure that the two interconnected technologies are fully optimized for each other. Industry working groups also have been created specifically to facilitate Open Cable network designs.
Open submarine networks
Building the best submarine cable is achieved by selecting the best wet plant and book-ended modems when and where required. But what about the rest of the network? Open, standards-based, software APIs allow for selecting and interconnecting the rest of the network components, and allow operators to implement the best software technologies related to network management, predictive analytics, closed-loop automation, real-time inventory, and more. Openness is now being considered across the entire network.
Coherent optical transmission modems
SLTE is an important area of technology innovation because it directly addresses ongoing bandwidth growth. New coherent modem transmission, intelligent power management, and spectrum sharing technologies are being introduced at a frenetic pace. The Shannon Limit looms large, but ways to side-step this physical limitation, such as SDM wet plants and its increased number of available fiber pairs, are being offered.
Packet networking evolution
The interconnection of terrestrial and submarine network segments—that is, the internet—is rapidly coalescing around a packet-based network, from access to data centers, and everything in between, including the submarine networks. Should (or will) packet switching and aggregation be integrated into SLTE, which is increasingly located directly within a data center? Common discussions in the submarine network industry are focused on how dispersed data centers and submarine networks can best interact to improve how future networks are designed, deployed, managed, and maintained. Given the fact that most of the new submarine cables being deployed and the overall traffic being turned up are related to submarine DCI of packet-based intra-data center networks, this topic will only increase in importance.
A sea change is on the horizon
This article addresses the most popular discussions expected to take place throughout 2020, some of which were started in 2019, albeit from more of a vision perspective. These technologies are rapidly changing from vision to reality, and are finding their way from R&D labs into real-world product networks. Many new technologies, originally targeted at terrestrial networks, are now finding their way into submarine networks. This is good news because the lines and demarcation points between submarine and terrestrial networks continue to blur, allowing for rapid innovation cycles to be applied to both network segments. These and other areas of innovation will continue to be discussed in 2020.