System requirements

Introduction

Before installing BIG-IP Next for Kubernetes (Host), make sure your environment meets the requirements in this section.

Important

  • BNK (Host) requires Red Hat OpenShift Container Platform with SR-IOV support.

  • The CPUs and SR-IOV virtual functions (VFs) allocated to TMM must share the same NUMA node.

Note

F5 does not validate the accuracy or safety of any third-party links listed below.

Supported host-node CPU architecture

The architecture of host-node CPUs must be x86-64 (sometimes referred to as AMD64). Both Intel and AMD host-node CPUs with this architecture are supported.

Minimum number of worker nodes

Your cluster must have at least three worker nodes so there are enough F5 Distributed Session State Management (dSSM) database replicas.

Supported MULTUS delegated CNIs

The following MULTUS delegated CNIs are supported:

  • macvlan

  • SR-IOV

Supported NICs for SR-IOV (x86-64)

For x86-64 hosts, when SR-IOV is used as the Multus delegated CNI, the following NICs are supported. When SR-IOV is not used, most NICs are supported, but for optimal performance SR-IOV is recommended.

Note

When using OpenShift, you must ensure that your version of OpenShift supports SR-IOV for a given NIC.

NIC

PF PCI ID

Linux driver

Min. kernel version

Min. NIC firmware version

Intel XXV710

8086:154c

i40e

5.14.0‑284.52.1.el9_2

9.1

Intel E810

8086:1889

ice

5.14.0-570.131.1.el9_6

4.4

Mellanox ConnectX‑5

15b3:1018

mlx5_core

4.18.0‑372.58.1.el8_6

16.32.3004

Mellanox ConnectX‑6 Dx

15b3:101f

mlx5_core

5.14.0‑284.52.1.el9_2

22.41.1000

Mellanox ConnectX‑7

15b3:1021

mlx5_core

5.14.0‑427.97.1.el9_4

28.47.1026

Mellanox ConnectX-8

15b3:1023

mlx5_core

5.14.0‑427.97.1.el9_4

40.47.1088

Broadcom BCM57414

14e4:16d7

bnxt_en

5.14.0‑427.97.1.el9_4

227.0.135.0/pkg 227.1.115.0

Broadcom BCM57508

14e4:1750

bnxt_en

5.14.0‑427.97.1.el9_4

226.0.145.1/pkg 226.1.107.1

Hardware requirements

Category

Component

Requirement

Notes

Server

Architecture

x86_64 minimum of 3 nodes

Enable SR‑IOV in BIOS.

Server

NICs

One physical SR‑IOV capable NIC

Recommended for separating internal and external networks.

Memory

Host Memory

Recommended minimum 256 GB

For application workloads and TMM.

Storage

Storage

Kubernetes persistent volumes

Required for F5 TODA Fluentd, F5 dSSM, and the TMM debug sidecar.

CPU

NUMA

CPUs and VFs on the same NUMA node

Set Performance Profile’s Topology Manager Policy to single-numa-node.

Software requirements

Software

Version

Notes

Red Hat OpenShift

See software support table

Evaluated versions mentioned above.

Cluster Network Operator

Included with OpenShift

Must be installed to manage cluster networks.

OVN‑Kubernetes CNI

Included with OpenShift

Default CNI for OpenShift. Calico (vanilla Kubernetes) also supported.

Multus

Included with OpenShift

Required for multi‑homed pods with SR‑IOV interfaces.

SR‑IOV Network Operator

Included with OpenShift

Manages VFs and network attachment definitions.

Helm

v3.x

Used to install BNK components.

Minimum Gateway API version

When installing a BNK (Host) instance on bare metal or a VM using F5 Lifecycle Operator, you’ll specify in the CNEInstance manifest whether you want to later configure the BNK instance using Gateway API (and the BNK Gateway API supplemental CRDs) or the BNK legacy service-proxy CRDs.

When using Gateway API, BNK requires Gateway API v1.4.1 experimental or later.

If the Gateway API CRDs

Then F5 Lifecycle Operator

Are not installed in the cluster

Will install an appropriate version

Are already present in the cluster

Will not alter those CRDs

Note

You must ensure the Gateway API CRDs are maintained at a version appropriate for BNK.

Pod networking

To support low‑latency 5G workloads, BNK relies on Single Root I/O Virtualization (SR‑IOV) and the OVN‑Kubernetes CNI. BNK also supports vanilla Kubernetes with Calico CNI. To ensure the cluster supports multi‑homed pods (the ability to select either the default virtual CNI or the SR‑IOV physical CNI), configure the following:

  • Cluster Network Operator — Must be installed to manage cluster networks.

  • SR‑IOV Network Operator — Manages SR‑IOV device plugin and network attachment definitions.

  • Multus — Enables pods to attach to multiple network interfaces.

Note

The Cluster Network Operator installation on GitHub.

CPU allocation

Multiprocessor servers divide memory and CPUs into multiple NUMA nodes, each having a non‑shared system bus. The CPUs and SR‑IOV VFs allocated to TMM must share the same NUMA node. Set the following parameters to ensure proper NUMA alignment:

Parameter

Value

Notes

Topology Manager Policy

single-numa-node

Set in the Performance Profile.

CPU Manager Policy

static

Set in the Kubelet configuration.

CPU Policy Option

full-pcpus-only

Ensures full physical cores are allocated to TMM.

Important

When full-pcpus-only is set with static CPU Manager policy, TMM receives both threads of the same core. No thread for the same core is assigned to other workloads.

Simultaneous multithreading (SMT)

BNK supports deployments in hyperthreading‑enabled environments, enhancing scalability and resource utilization. TMM can manage logical CPUs for high performance in hyperthreaded setups.

Persistent storage

The following components require an available Kubernetes persistent storage to bind to during installation:

Component

Notes

F5 TODA Fluentd

Logging collector

dSSM

Distributed session state database

TMM debug sidecar

Stores debug and diagnostic data.

Next steps