At CloudWorld this week, Oracle announced Generative Development (GenDev). At first glance, this appears to be yet another easy hook into the GenAI story. However, a closer look reveals this to be quite an interesting piece of new technology from Oracle. GenDev has the potential to make application development a lot easier, by looking at the underlying infrastructure in a fundamentally different way.
Application developers depend largely on the infrastructure on which they build the applications. If that infrastructure is very complex and data from multiple sources reaches the application in a complicated way, it means that the developer is not able to create the best application possible. With GenDev, Oracle wants to eliminate this concern as much as possible.
Juan Loaiza, EVP Mission-Critical Database Technologies at Oracle compares Oracle’s development of GenDev and its impact on application development to paving roads. “Just as paved roads had to be built for us to get the full benefit of cars, we have to change the application development infrastructure to get the full benefit of AI app generation,” he states. AI needs to be properly addressed and able to be picked up at the infrastructure layer, otherwise you end up with cobbled-together applications that often aren’t secure either, in part because it’s not clear exactly how they are put together.
Addressing data complexity at a different level
A central theme of GenDev is that Oracle is using this to solve the inherent complexity that data brings once you link it to an application on another level. Consider, for example, joining data from a JSON database and a relational database. GenDev handles this complexity at the data layer, so actually before a developer starts working with it. The idea then is that they don’t have to worry about the underlying infrastructure at all.
During a demo we saw, Loaiza showed some of what Oracle can do with GenDev. He gave the example of a developer who is going to develop an application that needs data from both a JSON database and a relational database. In doing so, the developer wants to work from the JSON perspective because files from such a document database are easier to read and control in the development process than data from a relational database.
JSON Relational Duality Views
The above may sound like something simple, but it is not. It means making data from a relational database available to developers on demand in JSON format. That’s no easy task, and this so-called JSON Relational Duality Views feature is something Oracle has been working on for a seriously long time, Loaiza admits during the demo.
JSON Relational Duality Views give data multiple personalities, in the positive sense of the term. It organizes data both relationally and hierarchically, at the same time. Oracle GenDev does all of this on the fly, without storing this new Duality View as such. You can create several of these Duality Views based on the same data from the same sources and let them coexist, by the way.
For developers, this means that they only need to link an application to a Duality View. They don’t have to worry about exactly what type of database the application addresses. It makes it possible to address the same data as a set of JSON documents and, at the same time, as a set of related tables and columns. Depending on the type of application, the connection can be made using APIs, such as the Oracle Database API for MongoDB and the Oracle REST Data Services, but it is also possible to use SQL or JSON functions for this purpose.
With JSON Relational Duality Views, Oracle is making the sometimes rather impenetrable relational databases a lot more accessible. In doing so, it is also capitalizing on the rising popularity of JSON databases, something it also did a few years ago by explicitly linking to MongoDB. If you want to know more details about JSON Relational Duality Views, Oracle’s website has a clear and easy-to-read document.
Enabled by AI in Oracle Database 23ai
JSON Relational Duality Views and other features of GenDev such as AI Vector Search obviously did not come out of the blue. GenDev would not have been possible without Oracle’s underlying database. Oracle announced Database 23ai earlier this year. This latest version includes all kinds of (Gen-)AI features that should make the application developer’s life easier. GenDev is built on top of some of these features.
With GenDev, it is possible to create modular applications without adding extreme complexity to the application structure itself. This should make applications more manageable, scalable, reliable, secure and generally more consistent. In other words, GenDev wants to make the benefits of AI available, while also minimizing its risks as much as possible. All of this is possible because of the underlying data engine, or Database 23ai, which takes care of all this before the application developer has to worry about it. In our opinion, Oracle GenDev is quite a fine piece of technology. Developers must surely appreciate what this can do for them. We expect at least a moderate amount of enthusiasm for those developers that look for a solution to their JSON/relational database issues.