The imputes for creating the tool is clear for AWS. Provide a way for customers to easily connect a SQL Server app to Aurora Postgres, saving big on licensing fees and reducing total cost of ownership. Assuming the tool is successful at some level, I’m sure it will provide a revenue boost for Amazon and some customers might (initially) feel a win. No harm, no foul on Amazon for leading the effort. Free markets, baby!
No matter how clever Babelfish is, however, I just can’t see how this is ultimately a win for SQL Server or PostgreSQL… or the developers that will ultimately need to support these “hybrid” apps.
I think Ryan makes good points and does hit upon the crux of the problem. I’d also say that there’s a secondary problem which Ryan hints at, but it is that a query may be sufficiently fast in one database variant but perform horridly in another. A classic example of this is a solution built on cursors in Oracle and then bringing that to T-SQL.