Load balancing should be possible with an equipment stack balancer or with programming. Programming arrangements have a tendency to be a reasonable piece less expensive. Applications would regularly associate with a database bunch by opening associations on one of the hubs keeping in mind the end goal to run exchanges. On the off chance that the database hub falls flat, the customer would need to reconnect to another database hub before it can keep on serving demands. Load balancing is regularly actualized as an additional layer: The application associates with the heap balancer by means of the system and the heap balancer interfaces with the database through the system. The heap adjusting can be executed on the application machines to dispose of the additional layer, this is better for inertness (each layer includes idleness) and as every application server has it’s own heap balancer it can likewise expel bottlenecks. The downside are that you’ll lose a solitary purpose of control and with various load balancers the quantity of wellbeing check questions will presumably increment. The two-layer setup is frequently utilized with Connector/J and haproxy.