F5 Expands Collaboration with NVIDIA to Turbocharge AI-Powered Edge Innovations
Over the past several years, service providers have been building out their infrastructures to prepare for the world of 5G. As 5G converges with the AI era in front of us, these investments are positioning telcos to optimize growth by providing accelerated AI services at the edge.
To help service providers achieve higher ROI on their infrastructures, F5 and NVIDIA have teamed up to accelerate F5 BIG-IP Next Cloud-Native Network Functions (CNFs) with NVIDIA BlueField-3 DPUs.
BIG-IP Next CNFs combined with NVIDIA BlueField-3 DPUs, enables customers to leverage F5’s proven network infrastructure capabilities—including edge firewall, DNS, policy enforcement and distributed denial-of-service (DDoS) refactored into lightweight cloud-native functions—while offloading and accelerating critical processes to the DPU to improve network performance and security. As a result, operators optimize their computing resources while gaining better power consumption per Gbps for edge AI services.
Our solution enables service providers and telecoms to manage demanding 5G and AI workloads—like AI-RAN—without overburdening CPU resources. By combining F5’s cloud-native traffic management and comprehensive security services with NVIDIA’s hardware acceleration capabilities, we’re helping our customers to reduce their total cost of ownership, accelerate the rollout of new services, and future-proof their infrastructures.
Leveraging underutilized RAN capacity for AI applications
Take AI-RAN, for example. At last year’s Mobile World Congress, NVIDIA joined with other telecoms, technology, and academic organizations to form the AI-RAN Alliance, with the goal of weaving AI into the fabric of the radio access network (RAN). Since then, NVIDIA has unveiled NVIDIA AI Aerial to serve as the building blocks for these efforts.
One of the opportunities is to leverage the underutilized compute capacity within 5G RAN infrastructures to run AI applications on the same network as existing RAN workloads. While doing so could open new revenue streams, it’s difficult to achieve without compromising latency and security. BIG-IP Next CNFs offloaded and accelerated on NVIDIA BlueField-3 DPUs simplifies this effort by delivering the necessary traffic management and security with the lowest latency for both 5G RAN and AI workloads.
Increasing resource efficiency at the edge
As more telcos move to a distributed UPF architecture to process data closer to the edge, they’re struggling to minimize space and power as they implement security and traffic management services on N6-LANs at these distributed UPF locations.
Our solution benefits telecoms by increasing the resource efficiency of these edge and far edge deployments. In the past, telcos used dedicated appliances or virtual machines for network functions such as firewall, DDoS, DNS, and traffic management, consuming the limited space and power typically available at edge locations.
Deploying BIG-IP Next CNFs on NVIDIA BlueField-3 DPUs enables service providers to significantly enhance operational efficiency while reducing cost. A notable example is a mobile provider that transitioned from virtual network functions (VNFs) to cloud-native network functions (CNFs), achieving a 31% reduction in vCPU usage. This shift translates into substantial cost savings for large mobile networks and unlocks additional capacity for high-value, revenue-generating telco services. By further offloading CNFs to BlueField, service providers can maximize vCPU availability, leading to even greater resource optimization and cost efficiency.
Addressing edge-related security gaps
As telcos increasingly deploy enterprise and AI applications at the edge to improve the customer experience and process data closer to the source, zero trust security becomes essential. However, a fragmented approach to network security can lead to blind spots and coverage gaps increasing the risk of vulnerabilities.
To ensure robust protection, telcos must adopt a comprehensive, end-to-end security strategy that eliminates these risks. By deploying BIG-IP Next CNFs on NVIDIA BlueField-3 DPUs, service providers can gain access to carrier-grade network functions, including F5’s firewall, DDoS protection, CGNAT and policy enforcement, in a single, cloud-native solution that leverages the zero-trust capabilities of BlueField. This integrated approach ensures robust, zero trust security for edge AI and 5G environments, enhancing both network protection and operational efficiency.
Unlocking the benefits of the AI era
This solution builds on previous collaborations over the past several months. In October, our two companies introduced BIG-IP Next for Kubernetes deployed on NVIDIA BlueField-3 DPUs, designed to help enterprises and service providers simplify the complexity of managing their large-scale AI factories and infrastructure.
This is just the beginning. Be sure to check back as we continue to develop the innovative solutions our customers need to thrive in the AI era.
F5 is sponsoring NVIDIA GTC as a silver sponsor, March 17-21, 2025, in San Jose, California. Learn more about F5 at NVIDIA GTC on our event landing page. Also, be sure to read our press release.