1. Support for Oracle-to-PostgreSQL migrations in Database Migration Service at Google Cloud Next ’23
2. Introduction of Gemini for Google Cloud at Google Cloud Next ’24, including Gemini in Databases feature
3. Preview of Gemini-assisted code and schema conversion for modernizing Oracle databases to PostgreSQL on Google Cloud, with assistive code explainability feature
At Google Cloud Next meetings in both ’23 and ’24, significant advancements in database migration and AI-assistive capabilities were announced. In ’23, Database Migration Service (DMS) introduced support for Oracle-to-PostgreSQL migrations with integrated code and schema conversions to streamline the process. This simplifies the most time-consuming phases of migration, making them more efficient and intuitive.
At Google Cloud Next ’24, Gemini for Google Cloud was unveiled, featuring AI-assistive capabilities based on Google’s Gemini models, including Gemini in Databases. The preview of Gemini in Database includes assistive code and schema conversion, enabling faster modernization of Oracle databases to PostgreSQL on Google Cloud. Additionally, Gemini-assisted code explainability simplifies code conversion, helping developers become adept at using the PostgreSQL dialect.
Migrating code and schema from Oracle PL/SQL to PostgreSQL can be challenging due to differences in syntax, data types, and procedural constructs. In Database Migration Service, Gemini assists with manual interventions by learning from a user’s edits and responses to real-time recommendations. This training enables Gemini to identify patterns and transformations in the code base, offering suggestions based on what it has learned.
For example, converting Oracle Bulk Collect to PostgreSQL’s equivalent involves manual adjustments in the DMS conversion workspace. Once this is done, Gemini can recognize similar occurrences and suggest further edits based on the user’s actions. This personalized assistance helps streamline the migration process and ensures a successful transition to Cloud SQL or AlloyDB for PostgreSQL.