With the virtualized and cloud-based transformation of telecommunication networks, the integration of 5G and IoT technologies, and the diversified development of industrial applications, the operation and maintenance of telecommunication networks are facing unprecedented challenges. Artificial Intelligence (AI) technology has natural advantages in solving problems such as high computing data analysis, inter-domain feature mining and dynamic strategy generation. It offers new models and capabilities of network operation and maintenance in the 5G era.
1. Ubiquitous AI Empowers Network Evolution
As a world’s leading telecom equipments vendor, ZTE Corporation has proposed the uSmartNet "Self-Evolution Network" solution to realize the intelligence of the entire 5G network with ubiquitous AI, empowering the network to fully evolve.
- One unified AI platform: The self-developed uSmartInsight AI platform provides all-round AI capability, supports mainstream Deep learning and machine learning framework, has built-in 100+ all-domain network AI algorithms, supports visualized zero programming AI model development, and makes scenario exploration easier.
- Three types of AI engines: AI engine, Light AI engine, and Real-time AI engine, which match the all layers requirements such as: Service Layer, O&M Layer and NE Layer. These AI engines are embedded in all 5G products by a modular form to meet the ubiquitous intelligent requirements of the network.
- 60+ AI Scenarios: Covering the full life cycle of planning, deployment, maintenance, optimization and operation, and helps operators to build efficient operation networks.
ZTE uSmartNet Self-Evolution Network Solution
In terms of architecture, the uSmartNet "Self-Evolution network" solution adopts the principle of hierarchical and closed loop to build an NE-level, Intra-domain, and Inter-domain intelligent network architecture. The AI capabilities are designed in a modular form at the NE Layer, O&M Layer and Service Layer, to build a self-evolution network with gradual evolution of capabilities and continuous superposition of value. Through network evolution, O&M evolution, and operation evolution, network operation and maintenance are simplified, helping operators to reduce OPEX and improve efficiency:
- Network evolution: The NE Layer introduces the Real-time AI engines, to generate dynamic policies based on real-time perception and analysis of service running status, perform resource scheduling and assurance as needed, and achieve the best match between network capabilities and user requirements.
- O&M evolution: The O&M Layer introduces the Light AI engines, to simplify the O&M of the network and reduce OPEX based on comprehensive information perception, analysis, and decision making. Make the machine gradually undertake human’s repetitive work, reduce the workload of O&M personnel, improve O&M efficiency, and reduce O&M costs.
- Operation evolution: The Service Layer introduces the AI engine, to build an intent-driven end-to-end intelligent closed loop, comprehensively improve the service capabilities for users, quickly respond to users’ needs and intent guarantee, and continue to increase operational income.
2. "Self-Evolution Network" Boosts Efficient Operation
In terms of application scenarios, the self-evolution network can provide operators with a series of network AI solutions in more than 60 scenarios, covering the full life cycle of network planning, deployment, maintenance, optimization, and operation.
- Network planning: Quickly identify valuable areas and user-demand-suppressed areas, and release suppressed traffic through fast capacity expansion and capacity balancing.
- Network deployment: The Automatic Integration Center (AIC) is an intelligent tool platform, which can realize intelligent design, one-click installation and deployment, and automated testing. The staff skills requirement was reduced by 1 level.
- Network maintenance: Through intelligent strategies and tools, we can reduce alarm work orders and improve the efficiency of fault handling.
- Network optimization: The LION intelligent network optimization solution has the characteristics of rapid self-evaluation, high-efficiency，self-decision, and precise self-optimization basically without human intervention. The base station energy-saving solution based on AI self-learning can perform network load and user prediction, flexible time configuration and differentiated thresholds. Base station energy consumption can reduce by more than 10%. Compared with traditional manual energy saving mode, the system KPIs are more stable and the users’ perception is better.
- Network operation: Actively identifies unsatisfied users and cells, provides guidance for active optimization, guarantees VIP care for VAP users, and achieves active O&M of user experience.
ZTE has conducted intelligent network verification in more than 20 commercial and pilot projects with global operators, which have achieved periodic results. In the field of network intelligence, ZTE will continue to invest in research and innovation to provide operators with all-round professional support for 5G network construction and intelligent network development, and to help operators to build efficient, green, energy-saving, secure and reliable intelligent networks.