Deploying autoscaling on GCP
You can deploy FortiGate virtual machines (VMs) to support autoscaling on Google Cloud Platform ( GCP).
Multiple FortiGate-VM instances can form an autoscaling group to provide highly efficient clustering at times of high workloads. FortiGate-VM instances scale out automatically according to predefined workload levels. This deployment achieves autoscaling by using FortiGate-native high availability (HA) features such as config-sync, which synchronizes operating system configurations across multiple FortiGate-VM instances at the time of scaleout events.
FortiGate autoscale for GCP is available for on-demand (pay as you go) instances in FortiOS 6.2.3 and later versions.
The standard deployment contains the following:
- Highly available architecture that spans two availability zones
- Virtual private cloud configured with public and private subnets
- Cloud NAT
- External-facing network load balancer (LB)
- Internal-facing network LB
- Cloud functions, which run Fortinet-provided scripts for running autoscaling. Functions are used to handle cluster creation and failover management.
- Firestore database which stores autoscaling configuration, such as primary and secondary IP addresses. Firestore is a nosql database hosted on GCP.
- Managed instance group and instance template.