Service providers are facing intense pressure to scale infrastructure for 5G, edge computing, and emerging AI workloads while maintaining the rigid performance and security standards required for carrier-grade services. To address this shift, F5 (NASDAQ: FFIV) is positioning its BIG-IP Cloud-Native Edition as a bridge for operators transitioning from traditional virtualized infrastructure to Kubernetes-based architectures. By extending core networking and security capabilities into containerized environments, the company aims to mitigate the operational complexities and visibility gaps that typically arise when moving critical network functions to distributed, cloud-native systems. This strategic move seeks to allow operators to leverage the agility of Kubernetes without sacrificing the granular control and stability essential for large-scale telecommunications and service provider environments.
F5 BIG-IP Cloud-Native Edition and eBPF Observability
F5 is introducing a suite of tools designed to integrate traditional application delivery and security into Kubernetes-native environments. The BIG-IP Cloud-Native Network Functions are intended to extend core services—including DNS, CGNAT, network security, subscriber policy enforcement, and traffic management—directly into these modern architectures. Complementing this, BIG-IP Next for Kubernetes operates at the Kubernetes network layer to provide essential networking and manageability. To solve the common problem of fragmented troubleshooting, the company is also launching BIG-IP eBPF Observability. This tool provides kernel-level telemetry, allowing network and security teams to examine application behavior, container activity, and network flows through a single, correlated source of truth.
The company is positioning these updates to support a broader range of protocols beyond standard HTTP-based applications. By offering Layer 4–7 support for TCP, UDP, HTTP/2, Diameter, SCTP, GTP, and SIP, F5 is enabling operators to migrate a more diverse set of network services into Kubernetes. Furthermore, the inclusion of subscriber-aware policy enforcement allows for the dynamic application of quality-of-service (QoS), traffic-shaping, and content-filtering based on specific service requirements. This approach is designed to help teams automate configuration and lifecycle tasks through Kubernetes-native workflows, facilitating a smoother transition from VM-based to containerized infrastructure.
Efficiency Gains in DNS and CGNAT Workloads
The technical viability of moving these services to Kubernetes is supported by independent testing conducted by The Tolly Group. The testing compared BIG-IP Cloud-Native Network Functions against traditional virtualized network function (VNF) deployments within the same Kubernetes environment. In DNS scenarios, the results indicated that F5 processed 1.9x more queries per second while utilizing 91.5% less CPU and achieving 48% lower average latency than the VNF configuration. These metrics suggest that containerized network functions can significantly increase workload density compared to their virtualized predecessors.
The efficiency gains were also evident in Carrier-Grade NAT (CGNAT) workloads. According to the Tolly Group report, F5 delivered comparable throughput to the VNF setup but required 79.7% less CPU and 98.5% less memory. Additionally, the testing showed an 82% reduction in p95 latency. By reducing the hardware footprint required for these critical services, F5 is suggesting that service providers can more efficiently manage compute and memory resources as network traffic and service demands scale. This capability is particularly relevant for operators managing the high-density requirements of 5G and edge computing, where infrastructure efficiency directly impacts operational costs and scalability.
Key Takeaways
- F5 BIG-IP Cloud-Native Edition provides Kubernetes-native support for protocols including TCP, UDP, HTTP/2, Diameter, SCTP, GTP, and SIP.
- Independent testing by The Tolly Group showed BIG-IP DNS workloads used 91.5% less CPU and processed 1.9x more queries per second than VNF deployments.
- The new BIG-IP eBPF Observability tool provides kernel-level telemetry to correlate application and network context across distributed Kubernetes clusters.
TechInsyte's Take
In our view, F5 is executing a defensive yet essential pivot to maintain its relevance in the era of cloud-native telecommunications. For years, the industry has relied on BIG-IP for physical and virtualized infrastructure, but the rapid migration toward Kubernetes threatened to leave traditional application delivery controllers behind. By embedding these capabilities directly into the container orchestration layer, F5 is attempting to prevent the "fragmentation of control" that often plagues service providers during modernization. The significant efficiency gains reported in DNS and CGNAT workloads are the most compelling aspect of this announcement; if these numbers hold at scale, they provide a powerful financial incentive for CTOs to accelerate their move away from heavy VM-based architectures. This isn't just about new features; it is about ensuring that the transition to 5G and AI-driven edge computing does not break the carrier-grade reliability that service providers demand.
Questions & Answers
How does BIG-IP eBPF Observability address the "siloed tools" problem in distributed environments?
It provides a consistent source of application and network telemetry by using kernel-level visibility. This allows network, security, and platform teams to examine container activity and network flows together, providing correlated context to determine if issues originate in the application, the container, or the network.
What specific protocol support does this update offer for non-HTTP workloads?
The enhanced traffic management includes broad Layer 4–7 support for a variety of protocols, specifically TCP, UDP, HTTP/2, Diameter, SCTP, GTP, and SIP, which allows operators to move more than just standard web applications into Kubernetes.
What were the documented resource savings for CGNAT workloads in independent testing?
According to The Tolly Group, the BIG-IP Cloud-Native Network Functions delivered comparable throughput to virtualized deployments while using 79.7% less CPU and 98.5% less memory.
Can operators apply specific policies to individual users within this architecture?
Yes, the solution includes subscriber-aware policy enforcement, which enables operators to dynamically apply quality-of-service (QoS), traffic-shaping, and content-filtering policies based on specific subscriber and service requirements.
Source: F5
