Viewpoint
With the accelerated development of digital transformation in the entire society, 5G has become an indispensable next-generation network infrastructure, which poses higher challenges to transport network O&M.
- The broad applications of 4G/5G hybrid networking, high-speed mobile broadband services and large-scale mobile IoT make the transport network scale and service capacity increase explosively. The network structure is complicated in many dimensions, and the complexity of O&M is beyond manpower.
- Due to diversified service application demands, the users have higher requirements for the quality and efficiency of network services.
- 5G needs a huge construction investment, and the overall revenue growth of the communications industry is slow, so network operators are facing great pressure to reduce Capex and Opex.
Faced with the challenges, ZTE believes that network automation is the inevitable trend of network evolution. Therefore, the automatic management and control integrated system for transport network is launched.
Features of Transport Management and Control Automation
As the brain of network automation and the backbone of the whole network ZENIC ONE (called UME in the China Mobile market), the ZTE’s transport management and control integrated system,has four features: full-lifecycle intelligent O&M, transport network full-scenario coverage, AI and openness.
The ZENIC ONE consists of three modules: Intention-driven engine, automatic control engine, and network awareness engine. They respectively implement service intention construction, network orchestration/management/control, and real-time network monitoring/analysis. It constructs a full-lifecycle O&M mode from planning & deployment and assurance to optimization to form an automatic O&M closed loop. The AI+BigData platform runs through the modules to provide auxiliary support for the automation of the management and control system, improving network service quality, simplifying the O&M process and reducing O&M complexity.
In the 4G phase, the NM does not have the management and control capability or cannot manage different scenarios at the same time, which waste resources and does not facilitate O&M. Developed on the basis of the microservice architecture, the ZENIC ONE uses one management and control system to cover all 4G/5G transport network scenarios, such as IP/IPRAN, OTN, PTN/SPN, IP+optical, network slicing and cloud-network convergence. This greatly saves resource overheads and improves user experience. .
AI further improves the automation level of the management and control system by empowering new capabilities to network O&M. Telemetry and iOAM technologies enable the real-time awareness of the network status at the NE layer. Therefore, the management and control system has the second-level data collection capability. The big data technology can be used to preprocess data and finally provide the high-quality data basis for analysis and reasoning. AI generates algorithm models for each scenario through training data, and further improves the capability of the management and control system in identifying and predicting potential risks such as network faults and network traffic through real-time analysis and prediction. With the continuous development and application of AI, the management and control layer software is gradually undertaking repeated work to provide a policy support for O&M personnel to determine the network status, effectively reducing the workload of O&M personnel, improving O&M efficiency, and continuously cutting O&M costs.
The openness of the transport network management and control system is divided into two aspects: One is the interaction with the ZTE’s internal upper-layer automation system to build 5G end-to-end O&M mode from wireless and transport networks to core network. The other is the open interaction with the upper-layer system of the customer. Through open APIs, customized development is carried out in accordance with user requirements to rapidly meet user demands.
Automatic Scenario of Transport Management and Control
- 5G BS power-on and automatic transport service provisioning
In the traditional O&M mode, the BS service provisioning mode is to transport first and then BS, and transport and wireless network coordinate with each other, resulting in low BS service provisioning efficiency. The management and control system allow automatic service provisioning after the BS powers on. The BS access is an action that triggers automatic service distribution on the transport side. The key prerequisite is how to distinguish BS access from non-BS access. The management and control system analyzes the performance data reported by access devices, and automatically identifies according to the intelligent algorithm whether the access is BS access. If yes, the management and control system automatically matches the proper data model in the planning database based on the port information of the access device, and delivers services to implement automatic service provisioning on the transport side, effectively improving the BS service provisioning efficiency.
- Intelligent Service Guarantee
Service guarantee includes multiple aspects, and the configuration of a new network is an important part. Traditional configuration involves huge workload and configuration error risks. If the network can be migrated, the Opex can be greatly reduced. The management and control system provides the intelligent configuration check function. It extracts the configuration feature structure of the device from the existing network to form the role fingerprint of the device. The fingerprint is an important parameter feature of a specific type of service. For example, the L3VPN service has parameters such as port, IP address and tunnel. The correct value range/value of the parameters of the L3VPN service in different scenarios constitutes the role fingerprint of the L3VPN service in this scenario. The principle of the function is as follows.
First, extract massive configuration data from the existing network devices, use the AI algorithm (e.g., NLP) to construct a hierarchical device model, and form a knowledge map of device configurations based on association rules at different levels.
Next, the role fingerprint is formed through cluster analysis. Then find the abnormal configurations by matching the role fingerprint corresponding to the device model when a new device is added or a new service is created on a device.
Finally, check the extracted abnormal configuration. The intelligent configuration check can shorten the configuration check period about 92%, and raises the fault identification rate to 85%, thus effectively reducing labor investment and potential network configuration risks.
- SDN Control Automation
The introduction of SR simplifies the control protocol and improves the network path adjustment and control capability of the management and control system. Path computation is a key module in the control field, and is widely used in scenarios such as re-routing and network tuning. Therefore, the PCE algorithm plays an important role in the performance of the management and control system. The ZENIC ONE integrates many self-developed PCE technologies, such as path packing, parallel resource deduction, and automatic stack depth optimization, to effectively improve the centralized control capability and network resource utilization of the system.
When a link fails, the management and control system guarantees the service SLA through BGP-LS-based dynamic rerouting to enable second-level recovery, the industry-leading performance and fast path switching. When faults occur at the same time in multiple locations of a network, the management and control system is aware of the network topology change in real time based on the BGP-LS, and judges according to the restoration arbitration mechanism whether to reroute. When a rerouting switching is triggered, the system calculates a new SR escape path in accordance with the policy and the self-developed PCE algorithm, and delivers the path to the device through the PCEP interface.
Network tuning includes link congestion re-optimization, global traffic balancing, latency optimization, and topology change re-optimization. The management and control system can analyze the collected data and determine the tuning target according to the requirements of the O&M personnel, and automatically generate the tuning policy through the self-developed iTE algorithm engine. After the determination by the O&M personnel, it delivers the device for network optimization, and feeds back and self-learns the tuning result to further improve the network tuning capability, forming a set of manual/automatic closed loop optimization system based on the core engine of the iTE algorithm.
- Network Slicing
Network slicing is one of the most distinctive features and advantages of 5G. Physical networks are logically divided into multiple virtual networks of different types to fulfill the requirements of different application scenarios. After the transport management and control system receives the user-oriented SLA parameters and application scenarios, the built-in TN NSSMF module obtains the planning data of the slice service to be activated according to the scenarios, and automatically acquire the resource information from the resource pool. The service management module dynamically allocates resource information according to the service configuration request and parameter requirements, and obtains data from the outside in the preset static mode or TOSCA model orchestration mode to supplement the configuration information and deliver the device construction slice. After a slice is created, the management and control system can monitor the slice dynamically, and guarantee the quality of the entire slice in real time.
At present, the ZENIC ONE system successfully puts into commercialization/trial commercialization in China Mobile, China Telecom, China Unicom, Velcom and other operators at home and abroad, and receives wide acclaim. In addition, ZTE works actively with operators to develop 18 innovative automation applications based on customer’s pain points this year. The development of the management and control system automation will gradually liberate people from the complicated network O&M, greatly reducing Opex while maximizing network potential. ZTE will continue to work closely with partners to promote continuous improvement of network automation capabilities, and ultimately build intent-driven and self-awareness networks.
