Two Nodes, Single/Multiple Data Centers with a Load Balancer

As in the first configuration, two Tellaros are deployed in one or two data centers, exposing the Tellaro's SOAP and REST web services on port 8181. The Tellaros in this scenario are behind a load balancer and the transactions can go to either Tellaro in the cluster; data is replicated asynchronously over ports 7001, 7002, and 7003.

In this configuration, application designers do not have to build in resilience to node or data center failures for their applications—the load balancer handles it for them. The applications send their web service requests to the load balancer's web service endpoint and the load balancer takes on the responsibility of managing the transaction relay to the appropriate Tellaro node based on the load balancer’s rules.

Failures of any single Tellaro are transparent to applications, since the load balancer recognizes the unavailability of the failed Tellaro and routes transactions to the available node. When the failed Tellaro rejoins the cluster, the load balancer recognizes this and starts routing transactions to the rejoined node again. The Tellaros synchronize their data automatically upon the clustered nodes seeing each other on the network again.