Huawei's IntelligentRAN 2.0 takes network automation to next level

Huawei has unveiled a major leap toward fully autonomous wireless networks with IntelligentRAN 2.0, an intelligence-powered architecture designed to minimize human intervention in network management to near zero.

The innovation arrives as telcos face operating pricing pressures due to rising network demands. While overall operating expenses in the industry saw a slight decline in 2023 — driven mainly by one-time divestments — labor costs surged by $7 billion and energy costs rose by 8%.

IntelligentRAN 2.0 tackles these challenges by using intelligent capabilities to automate costly maintenance operations. For instance, using digital engineers for routine inspections instead of deploying human technicians onsite or even detecting (and averting) network failures before they arise.

The announcement, made at last month's Global Mobile Broadband Forum 2024, builds on Huawei's 2022 release of IntelligentRAN 1.0, which marked a breakthrough in intelligent networks by introducing the Mobile Intelligent Engine (MIE).

The centerpiece of this latest upgrade is HANA (Highly Autonomous Network Architecture), which seeks to achieve "Level 4 network autonomy" — a state where networks can operate almost entirely without human involvement.

"Just as self-driving cars have evolved from simple parking assistance to full automation, we're transforming networks from systems that require constant human management to ones that can largely run themselves," said Calvin Zhao, president of Huawei Wireless MAE Product Line.

The system operates through a three-layer architecture. At the top service layer, which acts as the network's "brain" for high-level decision-making, Huawei has integrated its Telecom Foundation Model - a specialized AI model trained on over 30 years of Huawei's service support. This integration allows engineers to communicate with networks using natural language commands instead of complex technical inputs. This "flexible intent interface" enables operators to state their requirements simply, with AI translating these requests into specific network actions.

"Just as traditional cars require drivers to manually control every action, previous networks needed engineers to configure hundreds of specific technical commands," Zhao explained. "With HANA, network operators can simply state what they want — like telling a self-driving car your destination — and the system's AI automatically translates these requests into complex network actions."

HANA’s network layer functions like the "nervous system," coordinating and executing complex operations through its RAN Digital Twin System, which won a GSMA GLOMO award earlier this year. The digital twin creates a real-time, three-dimensional model of the physical network, enabling it to predict potential issues before they happen. This approach allows for testing network changes before implementation and lets multiple scenarios be analyzed and optimized at the same time.

At the device layer — functioning as the network's sensory system — HANA's digital sites can monitor the status of every antenna, tower, and dish, enabling automatic adjustments. This capability greatly lowers maintenance costs by reducing the need for physical site visits.

Together, these technologies enable what Huawei refers to as "scenario-tailored digital engineers," digital agents capable of handling complex tasks autonomously.

By leveraging the extensive data generated through the three-layer architecture, these intelligent assistants can continuously monitor, adjust and optimize network performance and energy consumption — tasks that once required teams of human engineers working around the clock.

The big question is, will IntelligentRAN 2.0 deliver?

Although telcos have been embedding AI into their networks to help cut costs for some time, the savings have been underwhelming, with few operators reporting specific reductions attributable to AI, according to Omdia.

Yet, Huawei is optimistic that Level 4 autonomous networks, powered by IntelligentRAN 2.0, represent a true advancement. For context, TM Forum reports that 84% of networks currently only achieve Level 1 or 2 autonomy. This low level of automation explains the limited cost savings achieved thus far and suggests substantial room for growth in network automation.

Recognizing this potential, Huawei plans to continue developing the technology in collaboration with the Global TD-LTE Initiative and other industry partners. It also recently launched the “Intelligent RAN, Ubiquitous AI Project” at Global Mobile Broadband Forum 2024.

"Huawei will continue to work with other industry players to explore innovative, intelligent applications and build a thriving ecosystem that promotes business growth," Zhao said.

Calvin Zhao, president of Huawei Wireless MAE Product Line.