BIG-IP Cloud Edition FAQ

The BIG-IQ is HA and if using VMWare, the tier 1 L4 traffic distribution is a BIG-IP HA pair (AWS uses ELB for tier 1). But the Per App VE is more of a stateless BIG-IP instance designed to pull full configuration from BIG-IQ as it's horizontally scaled out.

 

The service scaling group contains the requirements for auto-scale such as throughput, CPU, memory thresholds and will dynamically update the tier 1 traffic distribution of the new Per App VEs. In this case, a service scaling group would always start with 2 Per App VE's and maybe scale out to 4 or 5 depending on your requirements. Once the load normalized, it would then scale back to the original two Per App VEs.

 

Because it's dynamically creating BIG-IP's as needed, all redundancy is controlled by BIG-IQ's configurations and the tier 1 traffic distrubtion.

 

This is designed for those less-stateful applications that don't need the dedicated BIG-IP HA pairs deployed in the cloud (but BIG-IQ could manage both).

 

https://devcentral.f5.com/s/articles/application-auto-scaling-through-big-ip-cloud-edition-31517