BRIDGING JAVA OBJECTS AND RELATIONAL DATABASES

features

• The PersistenceBroker allows Java Programmers to store and retrieve Java Objects in/from JDBC-compliant RDBMS

• Transparent persistence: Persistent classes don't have to inherit from a persistent base class or to implement an interface.

• Persistence by reachability: All Objects associated to a persistent object by references are made persitent too.

• Mapping support for 1:1, 1:n and m:n associations.

• Automatic assignment of foreign key values.

• The Object / Relational mapping is defined in an XML Repository. The mapping is completely dynamic and can be manipulated at runtime for maximum flexibility

• Classes can be mapped to an arbitrary number of JDBC datasources.

• The Broker uses an internal ObjectCache which

  • maintains object identity (multiple lookups on a given table with same primary key will return one and the same object !)

  • reduces read and write overhead against databases

  • prevents circular reads in object read phase

  • is fully garbage collectable through usage of soft references (if objects are referenced by the ObjectCache but not from any Client-Application they can be reclaimed by the garbage collector)

  • The CacheMechanism is pluggable and can be easily replaced with user-defined implementations. There are also some predefined implementations available.

  • Provides a cache synchronization mechanism to avoid data corruption when using multiple OJB servers.

• Configurable Lazy Materialization through Proxy support in the PersistenceBroker. The user can implement specific Proxy classes or let OJB generate dynamic Proxies.

• Support for Polymorphism and ODMG Extents. You can use Interface-types and abstract classes as attribute types in your persistent classes. Queries are also aware of extents: A query against a baseclass or interface will return matches from derived classes, even if they are mapped to different DB-tables

• Support for Java Array- and Collection-attributes in persistent classes. The attribute-types can be Arrays, java.util.Collection or may be user defined collections that implement the interface ojb.broker.ManageableCollection.

• Sequence-Managing. The SequenceManager is aware of extents and maintains uniqueness of ids accross any number of tables. Sequence Numbering can be declared in the mappping repository.

• Reusing Prepared Statements, internal connection pooling.

• Supports BMP Entity Beans (see demo code in package test.ojb.ejb). The caching technique improve the performance of EJB applications.

• Integrates smoothly in controlled environments like EJB containers (wrt. SecurityManagers and JNDI based Datasources).

• ODMG compliant API, a Tutorial, and TestCases are included (see package test.ojb.odmg)
  Important notes:

• OQL is currently not fully implemented. (Aggregations and Method Invocations)

• ODMG implicit locking is partly implemented but does currently not maintain transaction isolation properly. To achieve safe transaction isolation client application must use explicit lock acquisition.

• Scalable client / server architecture that allows to build massively distributed systems.

• Distributed Lockmanagement supporting four pessimistic Transaction Isolation Levels (uncommited or "dirty" reads, commited reads, repeatable reads, serializable transactions).

• Optimistic locking support. Users may declare int or long fields as version attributes or java.sql.Timestamp fields as timestamp attributes.

• Cache synchronisation for distributed caches.

• Comes along with fully functional demo applications running against HSQLDB.

• Quality assurance taken seriously: More than 120 JUnit-TestCases for regression tests. JUnit tests integrated into the build scripts.

• Provides Log4J logging facilities.

• 100 %: pure Java, Open Source, LGPL

$FOOTER$