An organization has decided to deploy an active-active high-availability cluster in Google Cloud. Which three load balancing features are critical to…

In a FortiGate active-active high-availability (HA) deployment in Google Cloud Platform (GCP), the architecture relies on native cloud networking components.

1. Health Checks: GCP Load Balancers use health checks to monitor the status of each FortiGate instance. If an instance fails a health check, the load balancer automatically stops sending new traffic to it, which is the core mechanism for failover.

2. Session Affinity: Since FortiGate is a stateful firewall, all packets for a single session must be processed by the same instance. Session affinity (also known as session persistence) on the GCP Load Balancer ensures that traffic from a specific client is consistently directed to the same FortiGate, preserving session integrity.

3. Route Updates: For outbound and east-west traffic, one FortiGate is elected as primary. This primary instance uses the SDN connector to manipulate GCP route tables to direct traffic through itself. If this primary instance fails, a new primary is elected, which then makes API calls to update the GCP routes to point to itself, ensuring traffic flow is restored.

(Assuming other options based on common cloud networking features)

Global Load Balancing: This feature distributes traffic across multiple regions. It is not a critical requirement for a single-region active-active HA cluster, for which a regional load balancer is sufficient.

Content-Based Routing: This is a Layer 7 load balancing feature that routes traffic based on content like URL or HTTP headers. The fundamental HA mechanism operates at Layer 4 and does not require content inspection.

1. FortiGate for Google Cloud 7.6.0 Administration Guide, Section: "High Availability for FortiGate on GCP".

Health Checks & Session Affinity: The guide describes the use of GCP external and internal load balancers. The configuration of these load balancers requires setting up backend services with health checks to determine instance availability and session affinity to maintain statefulness. (Reference: "Deploying active-active HA for FortiGate-VM on GCP" chapter).

Route Updates: The guide explicitly states, "The primary FortiGate is responsible for programming the GCP routes (using the SDN connector) to direct outbound traffic and east-west traffic through itself... If the primary FortiGate fails, the secondary FortiGate becomes the new primary and programs the GCP routes to direct traffic through itself." (Reference: "HA for FortiGate-VM on GCP using external and internal load balancing > Failover" subsection).