Technique for addressing a cluster of network servers

Abstract

A single firewall or cluster of firewalls with a public IP address is interfaced to an internet public subnet to receive service requests for a cluster of network servers. A first private subnet with a plurality of private IP addresses is interfaced to the single firewall or cluster of firewalls to receive the service requests after passing through a firewall. A plurality of redundant load balancers with a respective plurality of private IP addresses are interfaced to the first private subnet to receive the service requests after passing through the first private subnet. The load balancers are interfaced to a second private subnet. The network servers with respective private IP addresses are interfaced to the second private subnet to receive the service requests from the load balancers. At an initialization time, a private IP address is defined for the network load balancer system within the internet access subnet. When one of the load balancers becomes primary at the initialization time or switches from a standby state to an active state, the network load balancer system private IP address is defined as an alias in an interface table to be recognized by the one load balancer. When the one network load balancer switches from the active state to a standby state, the network load balancer system private IP address previously defined as the alias is released from the interface table.