Looking into deploying F5 vCMP? First, take a look at my blog post on starting out with F5 vCMP concepts and then go through my step by step guide on F5 vCMP. You probably understand virtualization technology. F5’s version isn’t any different. The device is basically running a custom version of Linux KVM, with a bunch of custom drivers to enable things like hardware SSL acceleration for the vCMP guests.

Second, take a look at the configuration and implementation guide for F5 vCMP on their support page here. There’s a lot more detail in their guides and are obviously the best place to get F5’s recommendation.

Third, take time to plan your F5 vCMP environment. Prepare how much CPU and Memory resources each F5 vCMP guest will need. I’ve found that people tend to underestimate the resources F5 vCMP requires. They under allocate their guests, and then need to beef them up down the road. If you have the extra resources on the host, it won’t be a big deal, just a reboot of the guest. If you don’t have the resources, you’ll need to migrate the guest to another host, or buy another host. Also, it’s an expensive waste of resources to have one or two vCPU’s not used on the host, so plan appropriately and plan early!

The minimum size guest is going to have a single vCPU. Remember though, a vCPU is a hyperthread, which is only half of a CPU core. An F5 needs a lot of processing power, and unless you’re doing just about nothing on your guest, a single hyperthread is not going to be enough. Also, it means that you’re sharing your management plane activities with your traffic plane on that single hyperthread. Expect to have a very slow GUI when you have any decent amount of traffic going through the device.

Fourth, leave a comment, or contact us if you’d like to get some help specific to your environment!

Leave a Reply

Your email address will not be published. Required fields are marked *