Are there any plans to introduce the concept of automatic client
routing to the principal server in a cluster of N machines. For
example, if there is a four node replication cluster N1 .. N4, at any
time only one can be principal (the one which does the writing). In
Oracle and DB2, client side libraries provide a way for the clients to
connect to the principal writer regardless of where it is running on
N1 .. N4. This way client need to be aware of only one connection
string.
EnterpriseDb is a failover manager which relies on virtual IP
management, not the one I described above.