BEA hailed the Kodo engine for its performance and ease of use. It uses a WebLogic Workshop-like simplified visual approach to mapping object and relational entities, making persistence much easier for developers to create.
Additionally, BEA claims that SolarMetric customers are seeing run time performance boosts ranging from 10% to 1,000%, although it was unable to back up its figures or say which cases get the maximum benefit.
The deal was obviously conceived and executed very quickly, as BEA executives were not fully aware of the Kodo product’s capabilities.
But its plans are more definite. BEA will continue to offer Kodo as standalone product, so as to get opportunistic sales into WebSphere or Oracle shops. It will also incorporate Kodo into upcoming releases of WebLogic Server 9.x, and will support the standalone product on 8.x predecessors. Branding has yet to be determined.
BEA promises to blend JDO with its own implementation of EJB 3.0 as party of its blended technology: strategy that combines the best of all worlds, regardless of whether the technology is JCP-blessed. It is promoting the combination as the easiest, fastest path to developing Java applications.
Persistence has long proven a headache for enterprise Java applications because of its complexity. Not only does it involve mappings of data structures, but also the manipulation of related parameters such as caching. In traditional internal applications this was handled by the database engine. However, because the web is a stateless environment, persistence has often been the burden of middleware.
The Java community’s early response to persistence was container-managed persistence (CMP), a feature that proved one of the least popular elements of the J2EE specification because of its complexity. In recent years, alternate approaches such as Hibernate and TopLink have also emerged, fragmenting the Java community.
At this point, the Java community’s official plan for EJB 3.0 (also known as JSR 220) is to blend features from several persistence models, including Hibernate, which theoretically supersedes JDO. Although the JCP site denies that JDO will be killed, it adds that we expect over time JDO developers and vendors will shift their focus to the new persistence API.
For the record, BEA claims it’s not splintering technology off on a separate fork. BEA said it believes that JDO will, in fact, be incorporated into J2EE over time.
