Databricks Makes It Easier to Get Out of Oracle

Artificial intelligence that rewrites code and can reprogram old apps into newer formats is lowering the cost of switching enterprise software providers and is poised to help companies tackle previously nearly impossible tasks, as Aaron reported Monday. 

One such task we’re hearing more about involves companies getting their data out of Oracle, Microsoft SQL Server, Teradata and other databases that run in the companies’ private data centers. 

Some companies want to move data out of databases running in private data centers to the cloud so they can use AI from OpenAI or other providers to analyze their business or develop AI-powered applications.

Databricks and Snowflake, two cloud database firms, last month launched AI services that customers can use to get out of old databases.

Previously, this was time-consuming work that was handled manually by humans, said Arsalan Tavakoli, a Databricks co-founder.

The newfangled AI also helps ensure that after a database migration, employees can continue to use the same commands for handling the data that they previously used when it was in Oracle or other old databases, he said. Previously, companies hesitated to do a migration because such commands would change, not just because of the labor costs involved in the switch.

In the past, data-migration tools were based on pre-set rules that defined what types of data to move and how the data should be formatted. One key strength of new AI models is that they can translate variations in the commands that customers use to interact with different database products, known as “dialects,” so that older code can run on newer products, Tavakoli said.