In the era of 5G mobile Internet, higher requirements are put forward for optical transport network (OTN): larger bandwidth, longer haul, more flexible service types, higher quality and lower latency, and more flexible granularity-based scheduling. Optical network devices and management systems also evolve to be programmable, open, SDN-based, automated, and intelligent.
Physical restrictions on the optical transport layer should be considered for an automated, intelligent OTN. On the one hand, network provisioning involves service logic and physical parameters. In addition to accessible service routing, optical path indicators such as OSNR should meet the requirements to provide high bandwidth, long haul and high spectrum utilization while reducing relays and costs. On the other hand, environmental changes and the aging of optical components in network operation may lead to the deterioration of line quality, thus affecting service quality. Therefore, these factors must be taken into account in network maintenance in order to optimize and adjust automatically. 
The physical restrictions on the optical transport layer are one of the difficulties in OTN automation and intelligence. Through an in-depth understanding and research into the issues, ZTE has launched its adaptive optical network solution.

Overview
The adaptive optical network has such capabilities as beyond-100G software defined optics (SDO), automatic power optimization (APO), optical loss compensation, Flexgrid adjustment, and wavelength selective switching (WSS). It also supports AI, and can intelligently control programmable optical network operations based on the big data machine learning model to realize adaptive network O&M. 
SDO 
The beyond-100G OTN can choose different modulation modes and code patterns to meet the requirements of bandwidth, distance, and transmission quality in different application scenarios. ZTE's adaptive optical network has the SDO intelligence that can perceive network service status and transmission quality in real time, automatically select the appropriate rate, modulation mode and code pattern through an intelligent algorithm, and flexibly adjust the rate (100 Gbps to 800 Gbps) and set the modulation mode (PM-QPSK/PM-8/16/64/256QAM). The service signals of fixed rates and fixed modulation modes can also be slightly optimized and adjusted, so as to balance the transmission distance and spectrum bandwidth and achieve the best match between the transmission bandwidth and existing application requirements.
When the transmission distance is long, the service path deteriorates or traffic changes at the client, optimal adjustments can be made in time such as modifying the rate and modulation mode or shaping proper beams to improve transmission capacity. Operators can get a variety of solutions by investing in a single OTN board, and their networks are more adaptive to reduce the difficulty in network design and routine O&M and also lower the Capex and Opex.
APO 
In network operation, the changes of power parameters caused by network environment, human operation, and aging of optical components may affect the quality of services. ZTE's adaptive optical network has the APO function that can perceive the power status and changing trend of optical network in real time. If optical power parameters are degraded in the scenarios of network deployment, service scheduling, fault recovery, network adjustment, network optimization and long-term O&M, the adaptive intelligent control is used to balance the power of the multiplexing section and the channel layer and obtain a better optical signal-to-noise ratio (OSNR) before the service is damaged, so the impact of network changes on the service can be minimized. 

Architecture
ZTE's adaptive optical network consists of programmable adaptive optical network infrastructure and intelligent brain control layer (Fig. 1). The programmable adaptive optical network infrastructure is built in the OTN equipment such as ZXONE 19700/9700 and ZXMP M721, and the intelligent brain control layer in the management and control system ZENIC ONE. They communicate with each other via a NETCONF interface to intelligently control the adaptive optical network. 

Scenarios 
Through the automated optimization and on-demand modulation of the adaptive optical network, the spectrum utilization of the optical system can be significantly raised to get better system performance. More and more flexible collection of available path resources can also be provided to maximize transport and service recovery capacity and greatly improve OTN intelligence.
Constant Modulation Modes, Constant Transmission Rates
The scenario is applied to performance optimization of new or existing optical-layer services that have constant transmission rates, spectrum spacing and modulation modes. Through intelligent algorithms and APO, the adaptive optical network gains better OSNR and more WSS pass-through levels. When it predicts a degradation in the performance of optical-layer services, the following will occur:
—The priority is not to change the service route, and the optical power of optical-layer paths is automatically adjusted to compensate for degradation and restore the quality of service quickly. 
—When the performance deterioration is too serious and the automatic optical power adjustment cannot achieve the goal, the current services can be adjusted to other physical routes with better optical performance, and the optical power can be optimized automatically.
Variable Modulation Modes, Constant Transmission Rates
The scenario is applied to long-haul route recovery of new or existing optical-layer services. According to specific networking routes and optical fibers, the adaptive optical network uses AI learning models and intelligent algorithms to select policies, change modulation modes, or adopt the mixed modulation to improve transmission distance and performance. 
For example, the original 2×200G/16-QAM can be adjusted to 2×200G/8-QAM, which can extend the transmission distance and reduce the cost and complexity of electronic relay.
Variable Modulation Modes, Variable Transmission Rates 
The scenario is used for data center interconnect (DCI) over OTN that has variable modulation modes and transmission rates. According to periodic traffic changes between data centers, the adaptive optical network can intelligently and dynamically adjust transmission rates and modulation modes to achieve higher spectrum efficiency and consume less energy.    
The change of transmission rate from low to high as well as low power consumption are applicable to the scenario that requires short-term bandwidth growth, while the change of transmission rate from high to low suits the one with small traffic, where the reduced optical-layer rate at the line side leads to lower power consumption. For example, when the transmission rate at the line side goes down from 400 Gbps to 100 Gbps in the 400G/16QAM→100G/QPSK scenario, the power consumption of a single carrier may decline by 31% from 80W to 55W. 

Conclusion
Simplicity is the key to success. Based on the programmable optical network and intelligent management and control algorithms, ZTE has developed its adaptive optical network solution. The solution can be used to intelligently select and optimize optical network parameters in network deployment, adjustment and O&M, gradually replace manual operations and create a simple, efficient, flexible and open OTN for operators to cut their Opex and improve customer experience.