Performance Features

Caching Annotations

EclipseLink defines these entity caching annotations:

• @Cache

• @TimeOfDay

• @ExistenceChecking

EclipseLink also provides a number of persistence unit properties that you can specify to configure the EclipseLink cache (see "Persistence Property Extensions Reference" in Java Persistence API (JPA) Extensions Reference for EclipseLink). These properties might compliment or provide an alternative to the usage of annotations.

Using the @Cache Annotation

EclipseLink uses identity maps to cache objects in order to enhance performance, as well as maintain object identity. You can control the cache and its behavior by using the @Cache annotation in your entity classes. Example 18-1 shows how to implement this annotation.

@Entity
 @Table(name="EMPLOYEE")
 @Cache (
     type=CacheType.WEAK,
     isolated=false,
     expiry=600000,
     alwaysRefresh=true,
     disableHits=true,
     coordinationType=INVALIDATE_CHANGED_OBJECTS
     )
 public class Employee implements Serializable {
     ...
 }

For more information about object caching and using the @Cache annotation, see "@Cache" in the Java Persistence API (JPA) Extensions Reference for EclipseLink.

Read-only Queries

EclipseLink uses the eclipselink.read-only hint, QueryHint (@QueryHint) to retrieve read-only results back from a query. On nontransactional read operations, where the requested entity types are stored in the shared cache, you can request that the shared instance be returned instead of a detached copy.

For more information about read-only queries, see the documentation for the read-only hint in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Join Fetching

Join Fetching enhances performance by enabling the joining and reading of the related objects in the same query as the source object. Enable Join Fetching by using the @JoinFetch annotation, as shown in Example 18-2. This example shows how the @JoinFetch annotation specifies the Employee field managedEmployees.

@Entity
 public class Employee implements Serializable {
     ...
     @OneToMany(cascade=ALL, mappedBy="owner")
     @JoinFetch(value=OUTER)
     public Collection<Employee> getManagedEmployees() {
         return managedEmployees;
     }
     ...
 }

For more details on Join Fetching, see "@JoinFetch" in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Batch Reading

The eclipselink.batch hint supplies EclipseLink with batching information so subsequent queries of related objects can be optimized in batches instead of being retrieved one-by-one or in one large joined read. Batch reading is more efficient than joining because it avoids reading duplicate data. Batching is only allowed on queries that have a single object in their select clause.

Fetch Size

If you have large queries that return a large number of objects you can improve performance by reducing the number database hits required to satisfy the selection criteria. To do this, use the The eclipselink.jdbc.fetch-size hint. This hint specifies the number of rows that should be fetched from the database when more rows are required (depending on the JDBC driver support level). Most JDBC drivers default to a fetch size of 10, so if you are reading 1000 objects, increasing the fetch size to 256 can significantly reduce the time required to fetch the query's results. The optimal fetch size is not always obvious. Usually, a fetch size of one half or one quarter of the total expected result size is optimal. Note that if you are unsure of the result set size, incorrectly setting a fetch size too large or too small can decrease performance.

Pagination

Slow paging can result in significant application overhead; however, EclipseLink includes a variety of solutions for improving paging results; for example, you can:

• Configure the first and maximum number of rows to retrieve when executing a query.

• Perform a query on the database for all of the ID values that match the criteria and then use these values to retrieve specific sets.

• Configure EclipseLink to return a ScrollableCursor object from a query by using query hints. This returns a database cursor on the query's result set and allows the client to scroll through the results page by page.

For details on improving paging performance, see "How to use EclipseLink Pagination" in the EclipseLink online documentation, at:

http://wiki.eclipse.org/EclipseLink/Examples/JPA/Pagination#How_to_use_EclipseLink_Pagination

Cache Usage

EclipseLink uses a shared cache mechanism that is scoped to the entire persistence unit. When operations are completed in a particular persistence context, the results are merged back into the shared cache so that other persistence contexts can use them. This happens regardless of whether the entity manager and persistence context are created in Java SE or Java EE. Any entity persisted or removed using the entity manager will always be kept consistent with the cache.

You can specify how the query should interact with the EclipseLink cache by using the eclipselink.cache-usage hint. For more information, see "cache usage" in tJava Persistence API (JPA) Extensions Reference for EclipseLink.

Read-Only Objects

When you declare a class read-only, clones of that class are neither created nor merged greatly improving performance. You can declare a class as read-only within the context of a unit of work by using the addReadOnlyClass() method.

• To configure a read-only class for a single unit of work, specify that class as the argument to addReadOnlyClass():

  myUnitofWork.addReadOnlyClass(B.class);
  

• To configure multiple classes as read-only, add them to a vector and specify that vector as the argument to addReadOnlyClass():

  myUnitOfWork.addReadOnlyClasses(myVectorOfClasses);
  

For more information about using read-only objects to enhance performance, see "@ReadOnly" in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Weaving

Weaving is a technique of manipulating the byte-code of compiled Java classes. The EclipseLink JPA persistence provider uses weaving to enhance both JPA entities and Plain Old Java Object (POJO) classes for such things as lazy loading, change tracking, fetch groups, and internal optimizations.Weaving can be performed either dynamically at runtime, when entities are loaded, or statically at compile time by post-processing the entity .class files. By default, EclipseLink uses dynamic weaving whenever possible. This includes inside an Java EE 5/6 application server and in Java SE when the EclipseLink agent is configured. Dynamic weaving is recommended as it is easy to configure and does not require any changes to a project's build process

For details on how to use weaving to enhance application performance, see "weaving" in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Connection Pooling

Establishing a connection to a data source can be time-consuming, so reusing such connections in a connection pool can improve performance. EclipseLink uses connection pools to manage and share the connections used by server and client sessions. This feature reduces the number of connections required and allows your application to support many clients.

By default, EclipseLink sessions use internal connection pools. These pools allow you to optimize the creation of read connections for applications that read data only to display it and only infrequently modify data. The also allow you to use Workbench to configure the default (write) and read connection pools and to create additional connection pools for object identity or any other purpose.

In addition to internal connection pools, you can also configure EclipseLink to use any of these types of connection pools:

• External connection pools; you must use this type of connection pool to integrate with external transaction controller (JTA).

• Default (write) and read connection pools;

• Sequence connection pools; Use these types of pools when your application requires table sequencing (that is, non-native sequencing) and you are using an external transaction controller. Application-specific connection pools; These are connection pools that you can create and use for any application purpose, provided you are using internal EclipseLink connection pools in a session.

For more information about using connection pools with EclipseLink, see the following topics in EclipseLink Concepts:

• "Understanding Connections"

• "Understanding Connection Pools"

Parameterized SQL and Statement Caching

Parameterized SQL can prevent the overall length of an SQL query from exceeding the statement length limit that your JDBC driver or database server imposes. Using parameterized SQL along with prepared statement caching can improve performance by reducing the number of times the database SQL engine parses and prepares SQL for a frequently called query

By default, EclipseLink enables parameterized SQL but not prepared statement caching. You should enable statement caching either in EclipseLink when using an internal connection pool or in the data source when using an external connection pool and want to specify a statement cache size appropriate for your application.

To enable parameterized SQL, add this line to the persistence.xml file that is in the same path as your domain classes:

<property name="eclipselink.jdbc.bind-parameters" value="true"/>

To disable parameterized SQL, change value= to false.

EclipseLink determines binding behavior based on the database platform's support for binding. If the database does not support untyped parameter markers in specific SQL expressions, EclipseLink will disable parameter binding for the whole query.

To enable EclipseLink to restrict parameter binding decisions per expression, instead of per query, add this line to the persistence.xml file that is in the same path as your domain classes:

<property name="eclipselink.jdbc.allow-partial-bind-parameters" value="true"/>

For more information about using parameterized SQL and statement caching, see "jdbc.bind-parameters" in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Batch Writing

Heterogeneous batch writing is an optimization that allows EclipseLink to send multiple heterogeneous dynamic SQL statements to the database to be executed as a single batch. Batch writing is best used for applications that perform multiples writes in each transaction.

To configure batch writing, include the eclipselink.jdbc.batch-writing and eclipselink.jdbc.batch-writing.size properties in the persistence.xml file. The following example enables Oracle's native batch writing feature that is available with the Oracle JDBC driver and configures the batch size to 150 statements:

<property name="eclipselink.jdbc.batch-writing" value="Oracle-JDBC"/>
<property name="eclipselink.jdbc.batch-writing.size" value="150"/>

Different batch options are supported and custom batch implementations can also be used. For a detailed reference of the batch writing properties, see the batch-writing and batch-writing.size documentation in Java Persistence API (JPA) Extensions Reference for EclipseLink.

Serialized Object Policy

Serialized object policy is an optimization that allows EclipseLink to write out the whole entity object with its privately owned (and nested privately owned) entities and element collections into an additional field in the database. Serialized object policy optimizes fetching from the database, provides faster database reads, and reduces middle tier CPU and network access in certain situations.

Serialized object policy is best for read-only or read-mostly applications and should only be used for entities that load all their dependent entities or element collections. When using serialized object policy, database write operations (insert and update) are slower and queries for objects without private-owned data are slower. See "A Simple Serialized Object Policy Example" that demonstrates when serialized object policy is best used to increase performance.

Consider using serialized object policy only for complex objects with numerous aggregation as characterized by:

• Multiple database rows mapped to a single Java object

• When the object is read from the database all these rows are read at once (no indirection, or all indirection always triggered). There may be un-triggered indirection for other fields that are not included in the serialized object policy field

• If versionning is used, then updating or deleting any mapped row (or inserting of a new one) should result in incrementing of the object's version

• Object deletion causes all the rows to be deleted.

• Irregular structure of the aggregation makes it less possible to use other common optimizations (such as join fetching and batch reading).

Serialized Object Policy Configuration

Serialized object policy is enabled by using the @SerializedObject annotation on an entity or mapped superclass and passing in an implementation of the SerializedObjectPolicy interface. You must provide an implementation of this interface; there is no default implementation. The annotations also includes a field to define the column name for the object in the database. The default column name is SOP.

Example 18-4 enables serialized object policy, overrides the default column name, and sets optimistic locking to cascade, which can increase performance by keeping the serialized object policy field in the database up-to-date.

@Entity
@SerializedObject(MySerializedObjectPolicy.class)
@OptimisticLocking(cascade = true) 
public class Employee implements Serializable {
... 
 
@Entity
@SerializedObject(MySerializedObjectPolicy.class, column = @Column(name="ADDR_SOP"))
@OptimisticLocking(cascade = true)
public class Address implements Serializable {
...

Example 18-5 enables serialized object policy in the eclipselink-orm.xml file

<entity class="Employee">
    <optimistic-locking cascade="true">
    <serialized-object class="MySerializedObjectPolicy">
</entity>

<entity class="Address">
    <optimistic-locking cascade="true">
    <serialized-object class="MySerializedObjectPolicy">
        <column name="ADDR_SOP"/>
    </serialized-object>
</entity>

Example 18-6 enables serialized object policy in a customizer (either session or descriptor):

if (descriptor.hasSerializedObjectPolicy()) {

    MySerializedObjectPolicy sop = (MySerializedObjectPolicy)descriptor.
       getSerializedObjectPolicy();

    // to compare pk cached in SOP Object with pk read directly from the row from
    //pk field(s) (false by default):

    sop.setShouldVerifyPrimaryKey(true);

    // to NOT compare version cached in SOP Object with version read directly from
    // the row from version field (true by default):

    sop.setShouldVerifyVersion(false);

    // to define recoverable SOP (false by default):

    sop.setIsRecoverable(true);
}

To use a descriptor customizer, define the class and specify it using the @Customizer annotation:

public class MyDescriptorCustomizer implements
   org.eclipse.persistence.config.DescriptorCustomizer {
    public void customize(ClassDescriptor descriptor) throws Exception 
    {
         ...
    }
}
...
@Customizer(MyDescriptorCustomizer.class)
public class Employee implements Serializable {... 

To use a session customizer to reach all descriptors at once, specify it in a persistence unit property:

public class MySessionCustomizer implements
   org.eclipse.persistence.config.SessionCustomizer {
    public void customize(Session session) throws Exception 
    {
        for (ClassDescriptor descriptor : session.getDescriptors().values()) {
            ...        
        }
    }
}
 
<property name="eclipselink.session.customizer" value="MySessionCustomizer"/>

Read queries (including find and refresh) automatically use a serialized object if serialized object policy is enabled. If the serialized object column contains null, or an obsolete version of the object, then a query using a serialized object policy would either throw an exception or, if all other fields have been read as well, build the object using these fields (exactly as in the case where a serialized object policy is not used).

To disable querying the serialized object, set the SERIALIZED_OBJECT property to false as part of a query hint. For example:

Query query = em.createQuery("SELECT e FROM Employee e")
   .setHint(QueryHints.SERIALIZED_OBJECT, "false");

The following example demonstrates disabling searching for a serialized object:

Map hints = new HashMap();
hints.put("eclipselink.serialized-object", "false");
Employee emp = em.find(Employee.class, id, hints);

Applications that use serialized object policy should also consider using the result set access optimization. Use the optimization when querying to avoid the costly reading of the serialized object policy field (which can be large) if it is already cached and the query is not a refresh query. The optimization ensures that only the primary key is retrieved from the result set and only gets additional values if the cached object cannot be used. To enable the result set access optimization, set the eclipselink.jdbc.result-set-access-optimization persistent unit property to true in the persistence.xml file. For example:

<property name="eclipselink.jdbc.result-set-access-optimization" value="true"/>

A Simple Serialized Object Policy Example

Consider the following example object model:

@Entity(name="SOP_PartOrWhole")
@Table(name="SOP_PART_OR_WHOLE")
@Inheritance(strategy=InheritanceType.SINGLE_TABLE)
@Index(columnNames={"LEFTPART_ID", "RIGHTPART_ID"})
public abstract class PartOrWhole implements Serializable {
    @Id
    @GeneratedValue(strategy=GenerationType.TABLE)
    public long id;
 
    protected String description = "";
    
    @OneToOne(cascade=CascadeType.ALL, orphanRemoval=true)
    protected Part leftPart; 
    @OneToOne(cascade=CascadeType.ALL, orphanRemoval=true)
    protected Part rightPart;
}
 
@Entity(name="SOP_Whole")
@DiscriminatorValue("W")
@SerializedObject(MySerializedObjectPolicy.class)
@NamedQueries({
    @NamedQuery(name="findWhole", query="Select w from SOP_Whole w where w.id =
    :id", hints= @QueryHint(name="eclipselink.serialized-object", value="false")),
    @NamedQuery(name="findWholeSOP", query="Select w from SOP_Whole w where w.id =
    :id"),
})
public class Whole extends PartOrWhole {
}
 
@Entity(name="SOP_Part")
@DiscriminatorValue("P")
public class Part extends PartOrWhole {
}

The above data model allows the construction of a Whole object with any number of (nested) Part objects. For example:

• 1 level – A Whole object contains left and right Part objects (3 objects all together)

• 2 levels – A Whole object contains left and right Part objects; each of the Part objects has left and right Part objects (7 objects all together)

• 3 levels – A Whole object contains left and right Part object; each of the Part objects has a left and right Part objects; which each have a left and right Part objects (15 objects all together)

• n levels – (2n+1 - 1 objects all together)

Performance for the above data model increases as the number of levels in the model increases. For example:

• 1 level – performance is slower than without serialized object policy.

• 2 levels – performance is only slightly faster than without serialized object policy.

• 5 levels – performance is 7 times faster than without serialized object policy.

• 10 levels – performance is more than 25 times faster than without serialized object policy.