Improve RESTful response time - database

Question

I'm implementing a RESTful Service in Java EE with TomEE+(Apache CXF) which is using a database(specifically postgres). Now I've noticed that the longest time in my functions is spent in the getConnection() call for the database.
my code looks like this:
import java.sql.Connection;
import java.sql.SQLException;
import javax.annotation.Resource;
import javax.sql.DataSource;
import javax.ws.rs.Consumes;
import javax.ws.rs.Path;
import javax.ws.rs.GET;
import javax.ws.rs.Produces;
import javax.ws.rs.core.MediaType;
#Path("/test")
public class TestResource {
/* external DB resource
* configured in the resources.xml as "testDB". Either match the
* name or use the name parameter of the resource annotation.
*/
#Resource private DataSource testDB;
#Path("/hello")
#GET
#Produces("text/plain")
public String test() throws SQLException
{
Connection conn = testDB.getConnection(); //majority of time is spent in here
/*
do something.(e.g. PreparedStatement ps = conn.prepareStatement(...)
*/
conn.close();
return "world";
}
}
The database is defined in a resources.xml like so:
<?xml version="1.0" encoding="UTF-8"?>
<resources>
<Resource id="testDB" type="javax.sql.DataSource">
accessToUnderlyingConnectionAllowed = false
connectionProperties =
defaultAutoCommit = true
defaultReadOnly =
definition =
ignoreDefaultValues = false
initialSize = 0
jdbcDriver = org.postgresql.Driver
jdbcUrl = jdbc:postgresql://localhost/testdb
jtaManaged = true
maxActive = 100
maxIdle = 20
maxOpenPreparedStatements = 0
maxWaitTime = 100 millisecond
minEvictableIdleTime = 30 minutes
minIdle = 0
numTestsPerEvictionRun = 3
password = password
passwordCipher = PlainText
poolPreparedStatements = false
serviceId =
testOnBorrow = true
testOnReturn = false
testWhileIdle = false
timeBetweenEvictionRuns = -1 millisecond
userName = user
validationQuery = SELECT 1;
removeAbandoned = true
removeAbandonedTimeout = 60
logAbandoned = true
</Resource>
</resources>
So how could I reduce the time it takes to get a database connection? I'm already using the connection pooling mechanism. The only solution that comes to my mind is to make the resource class a singleton and get the connection once, but that seems counter-intuitive when many requests need to be worked on.

Answer 1

I would normally expect the interaction with the database to be the slowest part of the process. I would be interested in knowing how much overhead the database call is adding? How is that in relation to making the database call via a SQL tool outside of the Java code.
Also - I'm not sure if the example is just for brevity, but I would try to layer your solution so that the initial layer handles routing and validation, while the next layer handles any business logic and interacts with a database (DAO) layer.