Activating Zero Faults On Networks with SoftCOM AI - Huawei Certification

Competition drives innovation

In today’s telecoms landscape, IT and Internet companies are offering cloud services in traditional telco territory, including backbone networks, some MANs, IT infrastructure, and IT applications.

If operators cannot provide top-tier cloud services, they will lose many of their traditional services, in particular data center leased lines. Moreover, automation for O&M in telecom networks is low and network complexity means that 70% of network faults are due to human error.

Product innovation alone isn’t enough to overcome the challenges facing operators. Boosting competitiveness requires innovations in system architecture, products, and business models.

What is system architecture innovation?

In cloud computing, it isn’t a breakthrough in a server or storage product. Instead, it’s system-level innovation based on new distributed systems that increase resource utilization. Innovations in products, system architecture, and business models are mutually reinforcing.

On the product side, the principle behind Huawei’s network equipment design is high capacity and low latency. For system architecture, Huawei is looking at self-driving networks that are agile, automated, and smart.

What Do We Mean By Self-driving?

The aim of AI-driven autonomous networks is to create a self-driving network model with three features: agile devices, intelligent control, and intelligent analysis.

In telecom networks, the lower layer is network equipment and the upper layer is the control layer. For network-wide control and O&M, AI and segmented autonomous functions can achieve E2E functionality through the upper-layer operating system, thus enabling the entire network to become autonomous.

The biggest change realized by autonomous networks is that maintenance personnel are no longer involved in the entire service process. The entire network is self-driving in that it’s automated, self-optimizing, and self-healing.

Future networks will be fully automated at the operator side with Huawei providing fully automated AI-based online services. These services will be based on a continuous iterative AI model that’s available as a continually improving Model-as-a-Service.

Introducing AI to networks will bring new value from predictability. Telecom network management and the control center are based on device southbound interfaces and data collection. Various strategies and rules enable network-wide management and scheduling to fulfill three conditions for network automation: network reachability, SLA requirements, and resource efficiency.

However, as the network becomes increasingly complex, this isn’t enough. Online AI reasoning and data analysis are required to predict traffic, quality, and faults. Scheduling the network based on predictions of future conditions avoids faults before they occurs, optimizes quality before it deteriorates, and adjusts traffic before congestion occurs. Thus, the core value offered by AI is zero faults.

Five phases

Developing a self-driving network is a long-term process that we’ve divided into five phases:

One: AI knows “what happened.”

Two: AI can determine “why it happened.”

Three: AI can predict “what will happen” supported by manual judgments and decisions.

Four: AI judges “what measures need to be taken”, which are then carried out manually.

Five: Full automation enables self-healing.

Autonomous networks and Model-as-a-Service will provide end users with a minute-level  experience, optimal network connections at all times, and networks with zero downtime. Operators will benefit from a doubling in efficiency in three areas: O&M, resource efficiency, and energy efficiency.

Our experts say about Huawei Certification Exams