SDON: The Best Guarantee for Optical Network Reliability and Service Performance

The management and control system is capable of machine learning and has built-in optical optimization knowledge base, lab optical optimization supervised learning knowledge base, and online optical enhanced learning module, which can calibrate optical optimization policy in real time. Both the optical optimization knowledge base and the supervised learning knowledge base come from ZTE's huge online OTN data model base, and can be updated periodically to provide abundant learning samples for machine learning in the management and control system. The online optical enhanced learning module can calculate the optimal policy for the current system based on real-time network data.  

Typical Application Scenarios

The SDON solution enables the system to automatically deal with network emergencies and maintain service stability without human intervention. Its typical application scenarios are as follows: 

Optical Performance Degradation
 
Optical performance degradation includes aging of optical cable and substandard cleanliness of fiber pigtail joint. These factors cause optical cross-section power changes or line bit errors. The SDON solution can improve optical path performance and ensure reliable service transmission through APO and optical loss compensation while keeping line rates, modulation pattern and spectrum unchanged. 

The management and control system can monitor optical performance degradation and trigger optical optimization in real time. It enables the intelligent algorithm to match the optical optimization policy. The management and control components can generate a specific adjustment command set without changing the rate, code pattern, or spectrum. The management and control system delivers the adjustment command set to optical system equipment. The optical system components on the equipment optimize and adjust the optical power of service lines and compensate optical loss according to the command set. After the adjustment, they feed back the result to the management and control system. Based on the adjustment result and real-time service performance monitoring, the management and control system determines whether to perform the next optimization and adjustment. After the adjustments are completed, information about triggering reason, matching policy, and adjustment result will be put into the knowledge base, and the machine learning knowledge base will be updated iteratively. 

Under the condition of optical performance degradation, the SDON solution can achieve a better OSNR and pass through more WSS cascades by adjusting optical power or compensating optical loss without changing the rate, code pattern or spectrum. 

Optical Cable or Node Failure

When optical performance is degraded to a certain level, optical power optimization and optical loss compensation fail to reach the goal, or optical cable is interrupted or even the node fails, the management and control system will initiate the recovery policy. 

If there are paths with unchanged rates, code pattern and spectrum, the management and control system will use the algorithm and policies in the optimization policy base, calculate recovery routes, and select the optimal recovery path. After services are switched to the recovery path, the optical power is optimized to ensure optimal service performance.
If services cannot be recovered due to the limitation of physical resources such as devices and optical cables, the management and control system will first adjust the modulation code pattern under the condition of unchanged rates to ensure service transmission performance. If the modulation with unchanged rates cannot be implemented due to limited physical resources, the system will reduce transmission rates and match appropriate modulation code pattern and spectrum to ensure service connectivity.

Data Center Interconnect

As internet user traffic is bursty and unpredictable, it is difficult to completely fix the data center interconnect (DCI) model. If the OTN capacity is designed to reach the maximum traffic in a certain period of time, the bandwidth usage will be greatly reduced, and the network construction cost will be increased as well. SDON provides a solution to increase the transmission rate for DCI burst traffic. 

Traffic surge can be detected and reported to the management and control system, or predicted by the traffic prediction function of the management and control system. When network traffic increases sharply, the solution for increasing optical transmission rates will be triggered. According to the current performance of the optical system, the management and control system will calculate the optimal modulation code pattern and spectrum that can meet the transmission requirements, and generate a modulation command set. After receiving the modulation command set, the optical system components adjust the optical modulation code pattern and spectrum and increase transmission rates to ensure the transmission of burst traffic. They also adjust and optimize the optical power to ensure transmission performance. 

Whether in line performance degradation, optic cable interruption, node failure, or optical layer adjustment caused by traffic burst, information about triggering reason, adjustment policy, and optimization result will enter into the knowledge base for system machine learning, so that the algorithm can be optimized continuously.

Customer Value

ZTE's SDON solution can achieve a better OSNR through optical power optimization and optical loss compensation and guarantee service bandwidth and performance while keeping line rates, spectrum interval and modulation mode unchanged. In the case of limited resources, the solution can implement flexible spectrum modulation and line rate control, and improve the space for path selection. In the event of a failure, the solution can make full use of existing physical resources to ensure the service connectivity. The SDON solution can also automatically adjust the code pattern and rate to handle sudden increase in service bandwidth. 

With the accumulated optical data, growing SDON knowledge base, and rich optimization policies, the management and control system enhances its machine learning capabilities and can rapidly make the optimization policy that is most suitable for the current network and accurately predict the optimization results when an OTN network fails. It can be predicted that with the further development of AI, ZTE's SDON solution will provide the best guarantee for optical network reliability and customer service performance.