The F5 BIG-IP service (F5 BIG-IP Virtual Edition) provides intelligent L4-L7 load balancing and traffic management services at a local and global scale, robust network and web application firewall protection, and secure and federated application access. We can install more than one instance of this service as needed.
IBM Cloud for VMware Solutions offers promotions for some add-on services. Promotional pricing offers a number of months free of charge for a service's licenses, if the service has license charges. See Promotions for VMware Solutions add-on services.
The F5 BIG-IP service is not supported for vCenter Server with NSX-T instances. For vCenter Server with NSX-V instances, the installed version is v15.1.0.5.
Technical specifications for F5 BIG-IP
For information about resource requirements and capacity checking for some services, see Resource requirements for add-on services.
The following components are included with the F5 BIG-IP service:
Virtual machines
- Two virtual machines (VMs) with all options available.
- 2, 4, or 8 vCPUs per VM depending on the licensing option.
- 4, 8, or 16 GB RAM per VM depending on the licensing option.
Networking
- Private Virtual Extensible LAN (VXLAN) for High Availability (HA) synchronization.
- Access to Traffic Management Shell (TMSH) and Management Console through private management network.
Licenses and fees
License fees for each VM are applied to each billing cycle depending on the licensing option (Good, Better, or Best) and the selected bandwidth.
We cannot change the licensing level after service installation. To change the licensing level, you must delete the existing service and reinstall the service using a different licensing option.
Considerations when installing F5 BIG-IP
Before installing the F5 BIG-IP service, review the following considerations.
Based on the license model and bandwidth that you select, two BIG-IP VE VMs are deployed with the following configuration:
Maximum Bandwidth License Model: Good License Model: Better License Model: Best 25 Mbps 2 vCPU, 4 GB RAM 4 vCPU, 8 GB RAM 8 vCPU, 16 GB RAM 200 Mbps 2 vCPU, 4 GB RAM 4 vCPU, 8 GB RAM 8 vCPU, 16 GB RAM 1 Gbps 2 vCPU, 4 GB RAM 4 vCPU, 8 GB RAM 8 vCPU, 16 GB RAM 3 Gbps 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM 5 Gbps 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM 10 Gbps 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM 8 vCPU, 16 GB RAM
Additional considerations
- F5 BIG–IP limits the appliance throughput based on your chosen maximum bandwidth. Because network performance is affected by many factors, not all configurations and topologies may be able to achieve your chosen maximum bandwidth.
The pair of BIG-IP VE VMs suitable for high availability (HA) configuration will be deployed only into the default cluster.
In addition, 100% of CPU and RAM for the two BIG-IP VE VMs are also reserved because these VMs are in the data plane of the network communications and it is critical that resources are still available for them.
To calculate the CPU and RAM reservation for a single BIG-IP VE VM, use the following formula:
CPU reservation = CPU speed of ESXi server * number of vCPUs (from Table 1)
RAM reservation = RAM size (from Table 1)
IBM Cloud for VMware Solutions does not preconfigure HA. See the AskF5 article Creating an Active-Standby Configuration Using the Setup Utility.
- For information about the activities you should perform after you deploy F5 networks, see Working with F5 Networks BigIP in IBM Cloud for VMware.
Planning considerations
You must meet the following requirements to avoid failures with F5 BIG-IP:
- At least two active ESXi servers are available for the two BIG-IP VE VMs to be deployed with the anti-affinity rule of keeping the VMs on separate servers.
- The two active ESXi servers have enough resources available so that one BIG-IP VE VM can be hosted on each ESXi server with 100% CPU and RAM reservation.
- VMware vSphere HA has enough resources to host two BIG-IP VMs with 100% CPU and RAM.
Due to these requirements, you must plan for the space that is needed for F5 BIG-IP. If needed, before you order F5 BIG-IP, add 1-2 ESXi servers to the instance, or reduce vSphere HA CPU reservation for failover, or both.
F5 BIG-IP order example
You order a VMware vCenter Server Small instance with 2 ESXI servers with the following configuration: sixteen cores at 2.10 GHz each with 128 GB RAM. For F5 BIG-IP, you select the Best license model and a value of 5 Gbps for Maximum Bandwidth.
In this case, a single BIG-IP VM requires, on each server:
- 2.1 GHz * 8 vCPU = 16.8 GHz of CPU
- 16 GB RAM
In total, that is 33.6 GHz CPU and 32 GB RAM for two BIG-IP VMs.
Each ESXi server has a capacity of 16 cores * 2.1 GHz = 33.6 GHz, so we meet the first two requirements if both servers are active and there is at least 16.8 GHz of CPU and 16 GB RAM available on each server.
However, by default, vSphere HA reserves 50 percent of CPU and RAM for failover on vCenter Server instances that were initially deployed with 2 ESXi servers. For this example, the following is available:
50% of 2 * 16 cores * 2.1 GHz = 33.6 GHz available
There will be other workloads present on the ESXi servers, for example, VMware vCenter Server, VMware NSX Controller, and VMware NSX Edge. Because we are using these resources, we cannot satisfy the third requirement because we need 33.6 GHz of CPU and 32 GB RAM for the two BIG-IP VMs.
In this case, the F5 BIG-IP installation might fail, unless at least one ESXi server is added to the environment and vSphere HA failover reservations are updated appropriately to ensure that there are enough resources for two BIG-IP VE VMs. If additional resources are needed to run the F5 BIG-IP service, you can add more ESXi servers before installing F5 BIG-IP.
Considerations when you delete F5 BIG-IP
Before you delete the F5 BIG-IP service, ensure that the existing BIG-IP VE configuration is removed correctly. Specifically, network traffic must be routed around BIG-IP VE instead of through BIG-IP VE. Otherwise, the existing data traffic from your environment might be impacted.
Related links