Openness is an unstoppable juggernaut that is and will continue to penetrate all parts of the global network infrastructure, including the jugular veins that interconnect continental islands of connectivity – submarine cable networks. I already discussed the Advent of Open Submarine Networks last April, which stimulated much internal and external discussions as prognosticating about the future usually entails, and that’s a good thing.

The abovementioned blog discussed the coming of Open Cables, which has since come to commercial fruition with our partnership announcement last November with TE SubCom, an industry pioneer in submarine network technology. This partnership is a solid industry proof point of the first real step towards completely open submarine networks. I also touched upon multi-domain orchestration across different vendor terrestrial and submarine network solutions via standardized Application Programming Interfaces (APIs). So, is there anything else to discuss related to the benefits of open submarine networks? Yes, and the topic is big data analytics.

Application Programming Interfaces (APIs) are our friends

In the context of a multi-vendor end-to-end network service, say between two inland data centers interconnected by a submarine cable and terrestrial backhaul links on each end, APIs can be used for a variety of northbound and southbound use cases. Southbound use cases include configuring and managing network elements via a Multi-Domain Service Orchestration (MDSO) Network Management System (NMS), such as Blue Planet.

As long one has access to an open API, any active network element, such as modems, amplifiers, repeaters, branching units, and reconfigurable optical add/drop multiplexers, for example, can be virtualized and controlled by an NMS. This essentially turns submarine and terrestrial network segments into a single, seamless, and virtualized pool of network resources, some wet and some dry, that’s highly programmable. Of course, proper security and administrative mechanisms must (and surely will) be in place before this happens to ensure a very secure and reliable virtualized network is in place. Forward-looking submarine cable network operators can then more easily implement innovative new service offerings, such as Southern Cross’ elastic Gigaflex services, to best intercept the industry’s growing affinity for consumption-based service models. Times are indeed a changing!

This essentially turns submarine and terrestrial network segments into a single, seamless, and virtualized pool of network resources, some wet and some dry, that’s highly programmable.

From a northbound perspective, use cases such as alarms and network data extraction become a reality. Alarms can be generated and readily fed to an NMS and correlated across a multi-vendor end-to-end transoceanic network link allowing network operators to rapidly diagnose and effectively addresses inevitable issues that arise from time to time, such as a trawler having to go after diminishing fish stocks in much deeper waters where unprotected submarine cables lay, and snapping the cable causing an outage.

Terrestrial networks are more susceptible to faults, such as squirrel chewing on a piece of fiber – yes, a squirrel. There are some network elements of the end-to-end network, such as the Power Feed Equipment (PFE), which likely better suited from a safety perspective to only support northbound alarms. For example, would we want completely open PFE APIs for configuration given the high voltage it provides to the wet plant? I think not, but providing API-based alarms from such equipment, perhaps for shunt failures, is feasible.

Exactly what will be made available via open APIs will depend on a variety of factors related to security, safety, operational considerations, and much more, and this is normal and expected. It’s my belief that open APIs will be made much more readily available in one form or another.

Listen to your network

One of the biggest advantages of open northbound APIs is the ability to extract, consolidate, and analyze network data coming from the active (powered) network elements that support open APIs that offer a wide variety of timely network performance and health data such as polarization mode dispersion, chromatic dispersion, transmitter power levels, receiver power levels, losses, bit error rates, and so much more. In fact, much of this information is already being used to prevent, or at least minimize, the miserable days in the life of submarine field technician, as much of this test-related data can be readily and easily obtained from the submarine network itself anywhere in the world without having to travel or use 3rd party test sets, which we call GeoMesh Integrated Test Set Capabilities. How much information you collect and analyze will depend on what you have access to and what you want to do.

Big data analytics and machine learning algorithms have and will continue to change everything going forward.

What if you could poll all active network elements from your seamless terrestrial and submarine network, which we call a GeoMesh network here at Ciena, and then apply the very latest in big data analytic tools to turn oceans (pardon the pun) of data into actionable decisions? For example, wouldn’t it be amazing if your network told you that “repeater #33 located at these GPS coordinates has an 80% chance of failing within the next three months” or “modem #4 on the North American terrestrial backhaul link located at these GPS coordinates is exhibiting dribbling errors due to a bad splice” and then proactive address these impending outages before it occurs?

As my kids constantly question me, are we there yet? Well no, but it’s these types of network insights that are the ultimate end goal of big data network analytics – to listen to what your network is feeling and telling you so you can decide what to do, either proactively or reactively.

Isn’t this what our vehicles already do today with the “check engine light” and other dashboard indicators? Imagine if you told a mechanic of 50 years ago that in the future, the car itself would tell him that it’s detected a non-optimal mixture of air and fuel resulting from a faulty oxygen sensor, before he even opened the hood. He’d likely have thought you were breathing the exhaust fumes yourself, yet here we are today with self-diagnosing cars.

What I’ve come to realize over my career is to never underestimate what software developers can do once they gain access to APIs, data, and something they can control.

You could also monitor the overall health of your new network from cradle (RFS date) to grave (retirement date) and perform actions along its lifespan to ensure a highly reliable and available asset to end users that are so utterly dependent upon global network connectivity for a variety of reasons. Want 10 trillion of them? Check this out.


Will the global network infrastructure one day listen to itself, self-diagnose, self-heal, and self-optimize? Am I shooting too high for how far big data analytics can and will take us? Time will tell, but what I’ve come to realize over my career is to never underestimate what software developers can do once they gain access to APIs, data, and something they can control. The creativity of the software and network industry is truly astounding with a myriad of uncontestable proof points everywhere around us. Think about it, we used to go to the library to obtain knowledge, now the entire knowledge of mankind is essentially in the palm of our hands via our smartphones. Software developers have already had profound effects on other industries, and the global network is now on deck – batter up!

Join us at PTC 2017

Are you attending the Pacific Telecommunications Council 2017 event? If so, please join us at the “Convergence of Undersea and Terrestrial Networks – Where the Land and Sea Unite!” executive roundtable discussion in the Coral 1 Room on Tuesday, January 17, from 14:00 to 15:15 local time where this and other related topics will be discussed and debated. Ciena is also demoing a wide variety of innovative new technologies, including Big Data Analytics, so if you’re interested, please contact us for an appointment.