A new type of session bean that was introduced in Java EE 6 is the singleton session bean. A single instance of each singleton session bean exists in the application server.

Singleton session beans are useful to cache database data. Caching frequently-used data in a singleton session bean increases performance since it greatly minimizes trips to the database. The common pattern is to have a method in our bean decorated with the @PostConstruct annotation; in this method we retrieve the data we want to cache. Then we provide a setter method for the bean's clients to call. The following example illustrates this technique:

package net.ensode.javaeebook.singletonsession;  
 
import java.util.List;  
import javax.annotation.PostConstruct;  
import javax.ejb.Singleton;  
import javax.persistence.EntityManager;  
import javax.persistence.PersistenceContext;  
import javax.persistence.Query;  
import net.ensode.javaeebook.entity.UsStates;  
 
@Singleton  
public class SingletonSessionBean implements  
    SingletonSessionBeanRemote {  
 
  @PersistenceContext  
  private EntityManager entityManager;  
  private List<UsStates> stateList;  
 
  @PostConstruct  
  public void init() {  
    Query query = entityManager.createQuery(  
        "Select us from UsStates us");  
    stateList = query.getResultList();  
  }  
 
  @Override  
  public List<UsStates> getStateList() {  
    return stateList;  
  }  
}  

Since our bean is a singleton, all of its clients would access the same instance, avoiding having duplicate data in memory. Additionally, since it is a singleton, it is safe to have an instance variable, since all clients access the same instance of the bean.