Live cluster with two live Ozeki NG SMS Gateway Nodes "Setup 2"
In a live cluster setup both Ozeki NG SMS Gateway nodes are operational. One node serves one set of clients, the other node serves another set of clients. Both nodes have a set of active and a set of passive configurations. The passive configurations are in "Not connected" state. Passive configurations are defined for both: users and clients and service provider connections.
In this setup the primary system (Location 1) has a live accounts registered at the SMS service providers. These accounts are called "Account A1, Account A2, ..." on Figure 1. The secondary system (Location 2) connects to the primary system through an SMPP connaction. In this scenario all SMMS messages that are destined for the secondary system first arrive to the primary system. The primary system forwards these SMS messages to the secondary system.
During normal operation only the primary Ozeki NG SMS Gateway node is connected to the SMS service providers.
Normal operation
During normal operation, the load of the two nodes are balanced. Each node serves a configured set of clients. The primary node provides SMS service to the secondary node.
Failure of Node #2
If the secondary node goes down, the primary node can activate the passive configurations and can start to serve the clients of the secondary node. The primary node can safely detect that the secondary node is down, because the SMPP keepalive packate on account SMPP L2 will stop coming if the secondary node goes down.
Failure of Node #1
In case the primary node goes down, the secondary node can activate the SMS service provider connections normally used by the primary node. It can also activate the client configurations to serve the clients of the primary node. The secondary node can safely detect that primary node is down, because the SMPP SMS Service provider link provided by the primary node will become disfunctional.
Advantages
The advantage of this setup is that only one setup of SMS service provider accounts are required, and detection of the node failures is more straightforward. Automatic switchover is safer then Live cluster configuration 1
More information