How AMD is Helping Vodafone Meet AI Demands via its Chipsets

Share
The initiative addresses power consumption in mobile networks
AMD collaborates with leading telco Vodafone to develop mobile base station silicon chip designs, offering the chance for the telco to meet 5G AI demands

AMD has initiated a partnership with leading global telco Vodafone to develop AI-enhanced mobile network infrastructure at its Innovation Centre in Málaga, Spain.

Engineers from both companies are testing radio hardware equipped with AMD Zynq UltraScale+ RFSoC (Radio Frequency System-on-Chip) devices, single-chip platforms that combine radio features with Arm processors. These devices differ from traditional Application Specific Integrated Circuits (ASICs) through their ability to be modified remotely after installation.

“Vodafone’s goal is to develop a robust chip ecosystem.”

Vodafone, via announcement

The collaboration marks a shift in the development of telecommunications infrastructure, as network operators seek solutions to handle increasing data demands while reducing operational costs. Additionally, the partnership will focus on developing hardware that can adapt to changing network requirements without physical replacement.

Powering next-generation network architecture

Increased customer use of AI and cloud services across a wide range of key industries has also prompted networking expansion. With these new technologies continuing to be in high demand worldwide, data centres are continually working to expand their capabilities to keep pace.

Given the continued demand for data processing and communication services, 5G technology can help to improve operational efficiency within data centre facilities. Combining 5G and data centres can help businesses tap into low latency, high speeds and the large bandwidth that 5G can offer.

The partnership between Vodafone and AMD aims to reduce radio antenna size and energy consumption while maintaining network coverage. The AMD chips support Open Radio Access Networks (OpenRAN), a network architecture that enables hardware and software integration from multiple vendors.

Image: Vodafone

Such a design allows Vodafone to enhance network capacity during peak periods without replacing existing infrastructure. Engineers at the Málaga facility are evaluating compatible radio units from various manufacturers using the AMD UltraScale+ Zynq RFSoC DFE ZCU670 Evaluation Kit and AMD Vitis Design software.

This testing phase represents a significant step in telecommunications infrastructure development, as operators seek to reduce dependency on single-vendor solutions. The OpenRAN approach enables network operators to mix hardware and software components from different manufacturers, potentially reducing costs and increasing flexibility.

The implementation of AMD's RFSoC technology marks a departure from traditional radio unit design, incorporating digital signal processing capabilities directly into the radio frequency chain. This integration reduces the complexity of radio units while improving their adaptability to new standards and protocols.

“Vodafone’s goal is to develop a robust chip ecosystem by providing specifications and contributing towards the development of advanced algorithms that can power new energy-efficient and intelligent networks,” the telco said via its announcement.

Vodafone’s goal is to develop a robust chip ecosystem

Targeting energy efficiency

The initiative addresses power consumption in mobile networks, where signal amplification consumes significant energy. Vodafone is developing algorithms to match power consumption with customer demand variations.

Semiconductor integration will support advanced network features, including network slicing - a technology that enables dedicated network allocation for specific users with guaranteed service levels. This capability allows operators to partition their network resources efficiently, allocating bandwidth and processing power according to specific user requirements.

The testing process at the Málaga facility involves evaluating the performance of radio units from multiple vendors, ensuring compatibility with the AMD chipsets and Vodafone's network requirements. This multi-vendor approach aligns with the telecommunications industry's move toward more open and flexible network architectures.

Youtube Placeholder

Energy efficiency remains a central focus of the collaboration, with both companies working to optimise power consumption across the network infrastructure. The programmable nature of the AMD chips enables continuous refinement of power management strategies through software updates.

The partnership also explores the implementation of AI algorithms within the network infrastructure, enabling automated optimisation of network resources based on usage patterns and demand forecasts. This capability could enhance network performance while reducing operational costs.

With AI in mind, AMD has been continually expanding its high performance networking portfolio to address evolving system networking requirements for AI infrastructure. This has been to maximise CPU and GPU performance to deliver scalability and efficiency across entire systems.

"AMD is playing an important role in the advancement of 5G networks by providing the essential software and hardware for testing new Open RAN radio units," Vodafone says.


Make sure you check out the latest edition of Data Centre Magazine and also sign up to our global conference series - Data Centre LIVE ​​​​​​​


Data Centre Magazine is a BizClik brand

Share

Featured Articles

Vertiv & Compass Datacenters: Boosting Liquid Cooling for AI

Vertiv & Compass Datacenters develop combination liquid and air cooling systems to accelerate deployment of liquid cooling for data centre AI applications

Microsoft Unveils Zero-Water Cooling for AI Data Centres

Sustainability initiative promises elimination of cooling-related water consumption as Microsoft responds to growing water stress in key markets

How IBM's Optical Breakthrough Could Support AI Data Centres

IBM is developing co-packaged optics technology to enable faster chip communication, improved energy efficiency and longer data centre interconnect cables

How Evolution Data Centres & CREC Power Green Data Centres

Hyperscale

Lenovo: Data Centre Regulations to Transform Sustainability

Critical Environments

Re:Invent 2024: How AWS & Nvidia are Powering Ahead With AI

Technology & AI