5G is a network centered on customer experience, which needs to meet different scenarios and differentiated industry requirements in a timely manner. These requirements include not only large bandwidth, low latency and high security, but also efficient network construction and fast service provisioning, real-time network status awareness, fast fault diagnosis, precise traffic prediction and optimization, and system openness and reliability. Traditional telecom technologies, architecture and OAM models are difficult to meet these needs, so new concepts and technologies have to be introduced. With the industry's exploration in the fields of AI, big data, cloud computing and SDN&NFV in recent years, intelligence has been widely regarded as the core capability of 5G and future networks.
Based on a deep understanding of 5G wireline network needs and strong technical strength in the fields of SDN, machine learning, big data, knowledge graphs and intent networks, ZTE has innovated Athena 2.0, the intelligent network solution that allows for an intelligent closed-loop OAM throughout the lifecycle of a 5G wired network. The solution addresses the needs of network development and OAM in the 5G era and continues to evolve network intelligence based on advanced architecture to adapt to rapid network development and accelerate the arrival of autonomous networks.
Athena 2.0 Architecture
Athena 2.0 consists of two parts: a new-generation intelligent management and control system (ZENIC ONE) and a wired equipment network with super capabilities (Fig. 1). ZENIC ONE, the core of Athena 2.0, gives intelligence to the telecom network and functions as a network brain. The intelligent OAM is fulfilled at the management and control layer.
ZENIC ONE based on ZTE's self-developed cloud native platform contains three engines (intent engine, control engine, and awareness engine) and two platforms (BigData and AI). The three engines cooperate with each other to form an intelligent closed loop, which is supported by the two platforms.
The intent engine converts the user's intent input by voice and text into a network intent expression model and carries out scheme design and pre-verification. The control engine involves network orchestration, control and management service. It enables cross-domain and cross-vendor collaboration, offers fast end-to-end service distribution, and supports multiple networks such as IP, IPRAN, PTN, SPN and OTN. The awareness engine improves the ability and efficiency of network optimization through correlation analysis and in-depth mining of massive data, truly realizing end-to-end network optimization oriented to service and customer experience. With continuous optimization of user experience, it greatly enhances the ability of traffic optimization, early warning and problem prediction.
The BigData platform is the network data foundation of Athena 2.0 and also the twin data network of the physical network. It delivers rich data services at all levels such as structured and unstructured data services as well as knowledge graphs based on diagram database.
The AI platform is a basic platform that provides AI algorithms and services for various network services and components. It has a powerful AI framework and a variety of algorithms and interfaces, which are used for other parts of Athena 2.0 to improve its intelligence. It also makes continuous optimization based on the data on the BigData platform.
Intelligent Closed-Loop OAM
The Athena 2.0 solution enables intelligent closed-loop OAM throughout the lifecycle of a 5G wired network through rapid network construction, intent-based service provisioning, precise service awareness, automatic service recovery, intelligent fault diagnosis, and intelligent prediction optimization (Fig. 2).
Rapid Network Construction
In the traditional OAM, network construction or expansion requires manual and careful network design and planning, followed by site-by-site device activation, but there are the problems of low activation efficiency, high labor investment, and error-proneness. The function of rapid network construction can effectively solve the above pain points. First, it supports automatic deployment. After installation, the device automatically discovers network elements, boards and links and generates network topology without manual operations. Second, it globally plans logic resources such as routing domains, IP and network bandwidth, and forms different configuration templates for different scenarios. The user automatically creates configuration data based on the existing templates and sends them to devices in one click. The rapid network construction function reduces the complexity of site commissioning and improves the efficiency of network activation by 70%.
Intent-Based Service Provisioning
The traditional service provisioning requires users to configure tedious service parameters, which results in large workload, long provisioning time, error-prone operation and high OAM costs. The function of intent-based service provisioning is extremely simple to operate. The user only needs to select the service scenario and enter necessary information (service add/drop devices and ports), and then the system will automatically recommend the service-level agreement (SLA) data (bandwidth, protection attributes, and restoration attributes) according to user intents and history input information, with a friendly interface. After user selection, the system will automatically form multiple service solutions in line with the user's intent and give a recommended solution. Then the system converts the solution selected by the user into specific configuration information. The control engine verifies the chosen one and distributes it to related devices for service provisioning. The intent-based service provisioning is extremely simple to configure and visible in the entire process, which increases the efficiency of service provisioning by 80% and significantly enhances user experience.
Precise Service Awareness
The function of precise service awareness implements end-to-end real-time monitoring of various services. Inband OAM technology is used to collect statistics on quality data such as packet loss, delay and jitter of traffic streams and report them to ZENIC ONE in seconds through Telemetry. The BigData platform performs data storage and correlated processing, and the awareness engine fulfills realtime analysis and evaluation. If a certain quality feature of the service is found to exceed a threshold, the system will automatically trigger accurate monitoring, graphically show the monitoring results, pinpoint the deteriorated feature and give an early warning, so that the user can carry out targeted treatment, better the timely and proactive QoS management, and raise service guarantee capability and OAM efficiency.
Automatic Service Recovery
The function of automatic service recovery enables automatic protection and recovery at the service tunnel level. Based on the precise awareness of the awareness engine, the control engine executes recovery actions on a variety of tunnels. According to the information reported by the device, the control engine triggers the recovery module to intelligently calculate the routes and select a new route for the faulty tunnel. Even if ZENIC ONE is abnormal and unable to restore services, the device can automatically calculate the escape path based on the tunnel features and realize automatic service recovery. Athena 2.0 has powerful restoration to greatly enhance service reliability and satisfy SLA needs of high-value services.
Intelligent Fault Diagnosis
The function of intelligent fault diagnosis locates network faults on its own to significantly improve OAM efficiency. It comprises two parts. The first one is to form and update the network fault database. The BigData platform is responsible for data pre-processing of relevant fault data generated by the network, which involves data extraction, cleaning, and aggregation. On this basis, the AI platform executes the fault associated learning algorithm to create a series of related rules and update the results to the network knowledge base. In the second part, when the fault diagnosis is triggered, the system first identifies the root cause alarm based on the knowledge base rules, continues to locate the root cause of the fault based on the knowledge base rules, monitoring data, logs and other data, and offers the solutions. The intelligent fault diagnosis function records the effect of each execution and perfects the knowledge base to continually improve the efficiency and accuracy of fault diagnosis.
Intelligent Prediction Optimization
The function of traffic prediction analyzes and predicts the traffic of different regions and objects and identifies network bottlenecks as early as possible, so as to optimize or expand the network in time and ensure the long-term quality of network services. Traffic prediction contains two parts. The first one is to select the best prediction algorithm. The BigData platform pre-processes massive traffic data generated by the network, including feature extraction, classification and collection, to form data samples. The AI platform calls in multiple algorithms to predict the data samples and select the algorithm with the best predictive effect. In the second part, when the traffic prediction is triggered, the optimal algorithm is applied to the traffic data to be predicted to form a graphical prediction result and give a warning to the traffic that exceeds expectations. Similar to diagnosis, the traffic prediction function records the effect of each execution and continually optimizes the relevant algorithms through machine learning to further increase the efficiency and accuracy of traffic prediction.
Conclusion
Network intelligence has arrived as expected with 5G. Athena V2.0, ZTE's intelligent network solution, has been commercially used or trialed in networks of world-renowned operators such as China Mobile, China Unicom, China Telecom, Belarus A1, and Columbia TEF, and continues to play a role in improving user efficiency and experience. To actively adapt to the trend of the times, ZTE will work closely with global operators and partners to promote the development of intelligence and speed up the arrival of autonomous networks.
.png){kind=icon}
.png){kind=icon}
.png){kind=icon}
.png){kind=icon}