Several key factors of using SDN to realize network intelligent architecture

With the rapid development of globalization and information technology, network requirements are complex and changeable, network scales continue to expand, and network architectures are evolving rapidly. Some network management models will not be able to adapt to future network operation requirements. At the same time, the 5G network era centered on customer experience will accelerate the digital transformation of many industries. The network is facing diversified industry application services and network configuration challenges of ultra-large-scale connections, driving the transformation of telecommunications networks from a man-ruled model to an autonomous model, achieving intelligent Autonomous network.

With the rapid development of globalization and information technology, network requirements are complex and changeable, network scales continue to expand, and network architectures are evolving rapidly. Some network management models will not be able to adapt to future network operation requirements. At the same time, the 5G network era centered on customer experience will accelerate the digital transformation of many industries. The network is facing diversified industry application services and network configuration challenges of ultra-large-scale connections, driving the transformation of telecommunications networks from a man-ruled model to an autonomous model, achieving intelligent Autonomous network.

SDN network intelligent architecture

In December 2017, ETSI established the ZSM ISG industry standard group. The goal is to automate the entire process of network delivery, deployment, configuration, guarantee, and optimization. The ideal is 100% automation to help enterprises quickly transform to digital operations.

Based on the current situation and future development trends of telecommunications networks, ZTE proposes the overall intelligent solution for future networks, ZTE uSmartNet. As its sub-program, the bearer intelligent solution adopts SDN technology to provide automation support, and builds based on automatic control and willingness. The automated closed-loop network architecture built by the three engines of drive and network perception greatly improves the flexibility, agility and reliability of the network. The intelligent network architecture is shown in Figure 1.

Several key factors of using SDN to realize network intelligent architecture

Figure 1 Network intelligent architecture

The SDN-based intelligent network architecture is divided into three planes: the network forwarding plane, the management control plane, and the application plane.

● The network forwarding plane is used as an infrastructure to construct a simplified network, including the use of SR technology to simplify network protocols, improve equipment interface standardization capabilities and multi-vendor collaboration capabilities, thereby promoting end-to-end unified management that can be cross-domain, cross-layer, and cross-vendor.

● The management control plane introduces the three engines of “automatic control”, “will-driven”, and “network perception” to automate operation procedures and tasks through an intelligent automated closed-loop system, to ensure high fidelity of user wishes and network status, and to achieve “network freely” move”.

■ The automatic control engine realizes cross-domain and cross-layer network orchestration and control, automatically executes user wishes through the SDN controller, quickly completes automatic network deployment, uniformly dispatches and coordinates the entire network resources, and realizes automatic calculation of cross-domain and cross-layer network business routing , Automatic business issuance and automatic adjustment, so as to quickly achieve business goals.

■ Recognize business wishes with natural language or speech, convert wishes into precise network language and behaviors through the willing-driven engine, and support continuous network simulation verification, provide effective optimization methods, and ensure the multi-dimensional integrated intelligent management and control of the willing network.

■ The network perception engine realizes cross-domain and cross-layer network monitoring and analysis, and actively interacts with the network forwarding plane through the Telemetry interface. The system can collect network alarms, performance and logs in real time to ensure real-time awareness of the network status, and through the big data analysis platform, Perform dynamic analysis and real-time adjustments to the network.

● The application plane is responsible for building an application ecosystem for service providers, supporting ZTE and third-party APP stores for end customers, government and enterprise customers and partners to use, realizing mobile Internet customer experience, customers do not need to pay attention to the service layer, and quickly release and adjust anytime, anywhere Business, operation and maintenance are more convenient and simple;

The evolution goal and value of SDN network intelligence

The ultimate goal of network intelligence is to realize an autonomous network. The three engines drive network management to realize the intelligence of the design, configuration, change, and operation and maintenance of network lifecycle management, so that the network can be automatically controlled and operated automatically. system.

Network intelligence applies artificial intelligence deep learning and big data mining analysis technology to the operator’s network, replacing or optimizing the current manual work with intelligent subsystems, enabling operators to provide better network services more conveniently and efficiently , The value generated is reflected in the following aspects:

● Grasp the real-time operating status of the network through efficient big data collection and discover the changes in the network in time, so as to realize the intelligent analysis and responsibility positioning of efficient fault sources.

● Use AI technology to perform risk prediction and network assessment on user behavior, network services, and corresponding resource requirements based on historical data and real-time data.

● Match high-level operation wishes and strategies to achieve intelligent network management and control, deep perception and continuous verification and comprehensive optimization of these wishes.

● In-depth exploration of business opportunities through user wishes, combined with the company’s digital business needs and network pipelines, to provide users with continuous, agile, safe and growth guarantees for digital business, and to enhance the value of the network.

● Open network capabilities and APIs for management capabilities make services open, network programmable, and business customizable, building a complete end-to-end operation plan for users, and supporting DevOps tools to support online development, deployment, operation and maintenance, Cooperative operation supports the rapid customization and continuous delivery of innovative applications.

Typical application scenarios of SDN network intelligence

● IP+Optical Synergy Scenario

IP+Optical Synergy is based on the SDN architecture, with pipeline resource pooling. Through the hierarchical SDN controller, the joint planning and unified scheduling of the IP network and the optical network resource pool can be automatically realized. The pipeline resources can be dynamically adjusted in real time according to the traffic volume, and the network resource utilization rate can be from 20 %~30% increased to 60%~80%. Simplify operation and maintenance, support one-click service activation, one-time zero configuration activation from L0 to L3, and service activation time can be shortened to the minute level. In addition, IP+optical coordination also has many features such as topology on-demand construction, link on-demand bandwidth adjustment, multi-layer visibility of services, multi-layer cooperative protection, and survivability analysis.

Several key factors of using SDN to realize network intelligent architecture

Figure 2 IP+Optical Synergy

● Cloud and network integration management scenarios

Cloud-network integration management takes DC as the center and supports unified management of cloud access, cloud interconnection and other cloud-network collaboration. As shown in Figure 3, it supports the rapid opening of metro cloud dedicated lines, high-speed cloud interconnection configuration, and unified coordination of ZENICs in each DC. VDC controller supports end-to-end configuration of metropolitan area network and data center basic switching network. The ZENIC ONE cloud-network integration system supports intelligent DC deployment of value-added service network elements, and connects to VIM through the basic VNFM to realize the dynamic configuration of virtualized network elements in the DC and the automatic orchestration of the business chain.

Several key factors of using SDN to realize network intelligent architecture

Figure 3 Cloud-Network Convergence Management

The intelligent network operation and maintenance of cloud-network integration focuses on intelligent real-time network perception and fault location. Intelligent traffic collection based on Telemetry and traffic analysis based on big data and AI technology can automatically detect and analyze real-time traffic. It is required to detect and automatically optimize the deployment of traffic disturbances and traffic degradation before failure occurs. Realize automated closed-loop processing.

● Network tuning scenario

Network optimization can be classified according to different classification dimensions. From the network scope dimension, it can be divided: intra-domain network optimization, inter-domain network optimization; from the business scenario dimension: pure routing network optimization, MPLS network optimization; from the network level dimension can be divided : IP and light are independent of each other, and IP+light is optimized together. Each category will cross and contain each other, combined with network intelligent fault diagnosis and strategy center to form a closed loop of self-optimizing network, which can increase the reliability of the network by more than 80%.

Concluding remarks

The development of telecommunications network technology and network structure transformation have provided strong support for network intelligence, making network intelligence an inevitable trend in network development. Mainstream operators at home and abroad have already regarded network intelligence as a key capability for future digital operation transformation, and actively introduced AI capabilities to explore network intelligence models. Innovatively solve key issues such as business automation deployment, resource intelligent management and allocation, network fault intelligent analysis and positioning and prediction in telecommunications networks, and finally evolve to an autonomous network to achieve delivery, deployment, configuration, guarantee, and optimization of the entire process of intelligent operation. Realize the user’s willingness to deeply explore business opportunities, and combine the company’s digital business needs and network pipelines to provide users with continuous, agile, safe and growth guarantees for their digital business.

The Links:   AA084SB01-T2 AG320240A4

Leave a Reply

Your email address will not be published.