AWS Goes After Microsoft’s SQL Server with Babelfish for Aurora PostgreSQL

To get the best performance and proper lock ordering for concurrent updates I usually need to resort to using a variety of MSSQL only SQL extensions. (Lock hints, isolation levels, specific join methods, etc)

Is there a list of exactly what coverage this has? Does it support pinning execution plans?

Here’s an example of the crap I had to resort to in my last big project: https://github.com/realartists/shiphub-server/blob/master/Re...

I am seriously impressed if that actually works. For simply queries, sure, it's tricky, but it makes sense that you can do it. For complex queries, I have no idea how you'd even start to implement something like that.

Do they translate Transact SQL to Postgresql syntax, for do that build a new "language parser"-thing and use the internals of Postgresql to store and process the queries?

This sounds great but performance is the obvious question. Is this intended for long term usage? Or is this one of those stop gap solutions you're intended to replace but end up keeping for 5 years?

Also interesting to know how they map SQL Server security on top of this.

Very cool sounding but at the same time I might dread inheriting a system running on this if I started a new job.

This line cracked me up "AWS CEO Andy Jassy said. “Customers are fed up with and sick of incumbents.”"

Far more discussion yesterday, on the original AWS announcement: https://news.ycombinator.com/item?id=25267204

It seems quite interesting if it's 100% compatible. It has a lot of potential, but then customers may not want to migrates some tools since those tools don't officially support this product.

I wonder if this means that it will be compatible with SQL Server Management Studio (SSMS) and the SQL Data Tools (SDDT) that come with Visual Studio community edition.

One thing I really like about SQL Server aside from the high quality tooling is that I can write stored procedures which return multiple heterogeneous result sets with no ceremony. The TSQL dialect is one of the best IMO.

Won't this attract an API lawsuit?