This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Search
  • Building AI-boosted Intelligent IP Networks

    Building AI-boosted Intelligent IP Networks

After more than 30 years of development, IP networks have laid a solid foundation for network connectivity, and they’re critical to realizing the ubiquitous connectivity that will power an intelligent world. According to Huawei’s Global Industry Vision (GIV) 2025, 6.2 billion people will have access to the Internet and 100 billion connections will exist worldwide by 2025. Moreover, all enterprises will use cloud services and 85 percent of enterprise applications will be cloud-based. To achieve this, IP networks are required to carry more critical services, which in turn poses higher requirements on IP networks.

Detecting Usage Fluctuations and Diversity

A campus Wi-Fi network typically serves scenarios like office buildings, large stadiums, and large shopping malls. In these environments, the number of people in different areas fluctuates frequently and people use a variety of applications and services at the same time. O&M personnel have traditionally adjusted network resources manually. But to ensure service experience for different users, this approach is inefficient because it cannot cope with the rapid movement of people and assure user experience.

Can Wi-Fi networks become intelligent enough to detect fluctuations and multiple service types, and then automatically adjust resources to meet different service requirements? AI-boosted campus networks can intelligently detect changes in the number of terminals, access locations, bandwidth requirements, and service experience requirements of Wi-Fi users. It can also predict trends and dynamically adjust Wi-Fi network resources to optimize network performance.

Huawei has collaborated with leading customers to jointly develop Intelligent IP Networks, with test results showing that our AI-powered Wi-Fi solution can:

• Improve the throughput of wireless air interfaces by 58 percent over the industry average.

• Reduce Wi-Fi channel interference rate by 49 percent over the industry average. In addition, AI can be used for intelligent O&M on campus networks.

• Rectify about 85 percent of faults within 10 minutes.

What will Intelligent IP Networks Look Like?

Huawei believes that Intelligent IP Networks have the following characteristics:

• Super Capacity is the basis of the Intelligent IP Network architecture. At present, applications such as video, remote office, cloud computing, and AI are driving a new round of growth in network bandwidth. Campus networks are being upgraded with Wi-Fi 6 and 100GE switches, and data center networks and IP backbone networks are being upgraded to support 400GE. Advances in physical-layer performance and the emergence of network slicing networks like FlexE allow businesses to use bandwidths more flexibly and efficiently to simultaneously support multiple services, including office work, production, and computing, with a single physical network. Hard bandwidth isolation for traffic of different services enables 100 percent committed bandwidth to support verticals’ key services, businesses’ production networks, and operators’ IP private lines. The ability to intelligently adjust the bandwidth of different slices allows ultra-broadband networks to be flexibly adjusted for service changes, better addressing service needs.

• Intelligent Experience is the ultimate goal of an intelligent IP network. Many uncertainties exist with IP networks such as inadequate or non-existent negotiation on SLA between the service layer and the network layer. As a result, the expectations (the service intent) of the service layer are unclear at the network layer, creating uncertainty on the demand side. IP networks are statistically multiplexed, meaning that the resource usage level at the network layer constantly changes with service and traffic. That creates uncertainty on the supply side.

To eliminate these uncertainties, it’s necessary to accurately sense service intent. For example, the service layer could notify the network layer of service requirements, or the network layer could analyze service traffic characteristics (service models); proactively detect terminals, users, and service types; and infer the expectations of the service layer. These approaches can help eliminate demand-side uncertainty. Furthermore, a unified platform for network management, analysis, and control can use algorithms, such as neural networks, to establish network models, detect and analyze network status in real time, and learn about network resource usage. These capabilities help eliminate uncertainty on the supply side. Intelligent experience is also a process of matching service intent with network resources to continuously provide the desired connectivity services at minimum cost, thus achieving an application-driven experience.

• Autonomous driving is the key to improving user experience. Complaint-driven troubleshooting has brought significant challenges to network O&M, with the network O&M department often the last to know that a problem has occurred on the network. Proactive O&M is essential for improving user experience. First, network status should be monitored in real time to check whether an issue or potential risk exists on the network. If an issue or risk is discovered, AI can accurately identify the root cause by matching fault patterns and then automatically fix the fault before services and user experience are affected.

3-layer AI Architecture to Build Intelligent IP Networks

At HUAWEI CONNECT 2019, Huawei launched AI-boosted Intelligent IP Networks with three layers enhanced by AI:

• AI-boosted devices: Huawei provides a comprehensive range of AI Turbo products in NetEngine routers, CloudEngine switches, AirEngine WLAN products, and HiSecEngine security gateways products. These offerings deliver edge inference and real-time decision-making, and adjust IP packet forwarding policies based on service intent to ensure optimal service experience in real time.

• AI-boosted network management: Huawei iMaster NCE can identify the intent of the service layer, automatically generate and deploy network configurations, and ensure that the network meets service intent. It can also detect the health status of the physical network in real time, detect anomalies, provide alerts, and quickly offer handling suggestions. Its built-in expert system database enables the Huawei iMaster NCE to quickly troubleshoot and optimize against network anomalies. Huawei iMaster NCE also delivers real-time visibility of SLAs and enables predictive maintenance based on AI technologies. In addition, this system provides various viewgraphs of AI-powered network capabilities, enabling partners across various industries to perform customized development.

• Cloud-based AI Training: The Huawei iMaster NAIE comprises a cloud platform that provides a data lake, model and training capabilities, an open ecosystem, and developer services. The solution brings the following benefits: 1) It helps businesses develop AI algorithm experts and helps developers build AI algorithm capabilities. 2) It provides training services, so that developers don’t need to invest as much in computing power resources. 3) It provides a platform for sharing resource data that has undergone desensitization, which developers can use for model training. 4) It provides federated learning and transfer learning capabilities to tackle problems in model generalization and achieve model sharing.

AI training is the foundation of smart connectivity and smart O&M. In turn, building service, network, and fault models rely on training with big data and analytics. AI training can continuously evolve, enabling the entire system to become smarter, so that it adapts to rapid changes in services and networks to boost service quality and experience.

Practices and Experiences of Intelligent IP Networks

Intelligent IP networks not only vastly improve campus networks, they also deliver breakthroughs in Data Center Network (DCN), Wide Area Networks (WAN), and security firewall fields.

• DCN + AI: The arrival of the AI era poses higher requirements on DCNs. According to related tests, a packet loss rate of 0.1 percent in a DCN can reduce the computing power of AI training by 50 percent. To combat this problem, Huawei launched the industry’s first AI Fabric DCN solution, which achieves zero packet loss and fully unleashes the AI computing power on a DCN. This solution uses AI technologies to implement predictive traffic scheduling, achieving zero packet loss on the network and improving data computing and storage efficiency by approximately 30 percent. In addition, Huawei and leading customers have made great progress in joint innovation by applying AI technologies to autonomous driving of DCNs. Huawei’s solution can detect 75 types of frequent faults within one minute, locate them within three minutes, and rectify them within five minutes. Huawei’s AI-powered DCN solution can implement intelligence in understanding service intent, selecting the optimal network path, evaluating change risks, and detecting fault and the rapid location of fault root-cause. With these achievements, Huawei has taken the lead in creating an L3 autonomous driving network in the DCN field.

• WAN + AI: In today’s new era, a combination of 5G, cloud, and AI is powering all industries. 5G provides unprecedented capabilities for wireless access, while cloud and AI offer almost unlimited scalability for intelligent computing (for single tenants). The bonding between 5G, cloud, and AI—the DCN and WAN networks—shouldn’t be overlooked. The AI-powered DCN is the catalyst for adding AI to cloud, while the AI-powered WAN is the catalyst for joining the dots between 5G and cloud. We will use AI to advance autonomous driving networks in WAN networks and thus unleash the full potential of 5G, cloud, and AI, enabling millions of enterprises to migrate to cloud and bringing the benefits of 5G to all industries.

So, how can we make this reality?

Much like the DCN scenario, WAN networks can use AI to develop autonomous driving networks. Specifically, the AI-powered WAN can intelligently match network resources and intelligently select the optimal routes based on SLA requirements such as service latency. However, unlike the DCN scenario, WAN networks need to resolve how to quickly provision WAN networks to meet the different SLA requirements of various industries, for example, 5G telemedicine, where E2E latency must be less than 15 ms. Enabling the physical forwarding plane “body” to keep pace with the AI-powered “brain” for management, control, and analysis is a new challenge for WAN networks. Millions of enterprises are now migrating to cloud. Traditional WAN networks need to be manually provisioned hop by hop and so deployment efficiency is very poor. As virtual machines and containers can be provisioned much faster, WAN network deployment is the bottleneck. The source routing mechanism of Segment Routing IPv6 (SRv6), a next-generation routing protocol, shifts away from traditional E2E, hop-by-hop provisioning to source node provisioning only. SRv6 greatly simplifies WAN deployment and enables the body to keep up with the brain, realizing automatic and fast deployment in WAN networks.

5G-powered industries have varied SLAs, especially in terms of latency requirements. To address this, the WAN uses the SRv6 protocol to program the network forwarding route based on the optimal path calculated by the management, control, and analysis system. A route with a deterministic node, route, and latency can be quickly configured to meet the requirements of the service layer.

Therefore, SRv6 is a crucial forwarding plane capability of next-generation AI-powered WAN networks. SRv6 enables the WAN to intelligently recommend the optimal route, quickly deploy the optimal connections, and optimize service SLAs in real time. Together with 5G and cloud technologies, SRv6 can enable millions of enterprises to move to cloud.

• Network security firewall + AI: Malware has many variants and is difficult to detect, especially by today’s firewalls that use signature matching. Huawei confirmed its leadership in the industry by launching the industry’s first T-level AI firewall series, HiSecEngine USG12000. It handles threats that traditional firewalls cannot detect and uses a unique threat detection AI Engine (AIE) to identify, for example, compromised hosts and communication with external Command and Control (C&C) servers—at network borders in real time. Achieving a detection accuracy of more than 99 percent and powered by the AI chip, HiSecEngine USG12000 improves threat detection performance five-fold. By applying intelligent security event analysis and intelligent security policy optimization technologies, HiSecEngine USG12000 achieves service rollout in minutes and implements service-driven policy deployment and change, reducing OPEX for security O&M by 80 percent. The next-generation AI firewall will provide intelligent network border protection and build impenetrable high security for enterprises.

Summary

Customer-centricity is Huawei’s core philosophy. Customer needs are always the driving force behind Huawei’s development. Through the NetCity joint innovation program, Huawei combines the requirements of leading customers with its own R&D capabilities to develop leading IP network solutions and shape the future of IP networks with its influence in the IP standards community. Huawei will continue to work with customers and partners worldwide to continuously incubate cutting-edge products and solutions and lead the way in intelligent IP networks.

TOP