Fortinet FortiGate 400, Installation & Configuration Manual, Page: 76 / 308

76
Fortinet Inc.
Active-active HA High availability
During startup the members of an HA cluster negotiate to select the primary unit. The
primary unit allows other FortiGate units to join the HA cluster as subordinate units
and assigns each subordinate unit a priority.
The primary FortiGate unit sends session messages to the subordinate units through
the FortiGate HA interfaces. All FortiGate units in the cluster maintain all session
information. If the primary FortiGate unit fails, the subordinate units negotiate to select
a new primary unit. All connections are resumed by the new primary unit.
During a fail-over, the new primary unit notifies the adjacent networking devices so
that the entire network can quickly converge to the new data path. The new primary
unit also alerts administrators of the changes to the HA cluster by writing a message
to its event log, sending an SNMP trap (if SNMP is enabled), and sending an alert
email.
If a subordinate FortiGate unit fails, the primary FortiGate unit writes a message to its
event log, and sends an SNMP trap and an alert email. The primary FortiGate unit
also adjusts the priority of each of the remaining units in the HA cluster.
Active-active HA
Active-active (A-A) HA provides load balancing between all of the FortiGate units in an
HA cluster. An active-active HA cluster consists of a primary FortiGate unit and one or
more subordinate FortiGate units all processing traffic. The primary FortiGate unit
uses a load balancing algorithm to distribute sessions to all of the FortiGate units in
the HA cluster.
In active-active HA mode the primary unit uses one of the following scheduling
algorithms to distribute network sessions among the members of the HA cluster.
Table 15: Active-Active HA load balancing scheduling algorithms
Schedule Description
None No load balancing. Used when the cluster interfaces are connected to
load balancing switches.
Hub Used when the cluster interfaces are connected to a hub. Distributes
traffic to units in the cluster based on the Source IP and Destination IP
of the packet.
Least-Connection Distributes traffic to the cluster unit with the fewest concurrent
connections.
Round Robin Distributes traffic to the next available cluster unit.
Weighted Round
Robin Similar to Round Robin but weighted values are assigned to each of
the units in a cluster based on their capacity. For example, the primary
unit should have a lower weighted value because it handles scheduling
and forwards traffic.
Random Randomly distributes traffic to cluster units.
IP Distributes traffic to units in a cluster based on the Source IP and
Destination IP of the packet.
IP Port Distributes traffic to units in a cluster based on the Source IP, Source
Port, Destination IP, and Destination port of the packet.