HA-JDBC is a JDBC proxy that enables a Java application to transparently access a cluster of identical databases through the JDBC API.
HA-JDBC has the following advantages over normal JDBC:
The database cluster is available to service requests so long as at least one database node is active.
- Fault Tolerance
Because HA-JDBC operates via the JDBC API, it is transaction-aware and can survive a database node failure without failing or corrupting current transactions.
By balancing read requests across databases, HA-JDBC can meet increasing load by scaling horizontally (i.e. adding database nodes).
(source & for more info: http://ha-jdbc.github.io/doc.html)