- F5 (FFIV, Financial) collaborates with NVIDIA to enhance AI application delivery with new features for F5 BIG-IP Next for Kubernetes.
- Sesterce validates the integration, reporting a 20% improvement in GPU utilization.
- The solution introduces enhanced performance and security measures, optimizing multi-tenancy for large-scale AI infrastructures.
F5, Inc. (FFIV), a leader in application and API security, has announced advancements in its F5 BIG-IP Next for Kubernetes, integrating with NVIDIA's BlueField-3 DPUs and DOCA software framework. These enhancements, validated by European infrastructure operator Sesterce, aim at optimizing AI-first application delivery.
The integration promises a 20% improvement in GPU utilization by enhancing performance, multi-tenancy, and security. Key features include NVIDIA Dynamo and KV Cache Manager integrations, reducing latency for large language model inference systems. Additionally, smart LLM routing on BlueField DPUs efficiently manages workloads requiring multiple models, contributing to enhanced customer experiences.
Further, the solution includes scaling and securing Model Context Protocol (MCP) with reverse proxy capabilities and robust F5 iRules for rapid customization, supporting evolving AI applications and security requirements. This collaboration underlines F5 and NVIDIA's commitment to delivering high-performance traffic management and security for AI infrastructures.
The F5 BIG-IP Next for Kubernetes, bolstered by NVIDIA technology, offers enterprises improved control over traffic routing to AI factories, optimizing GPU efficiency. Available now, this solution represents significant progress in AI infrastructure support, as F5 continues to innovate alongside NVIDIA for scalable and secure enterprise AI solutions.
