Which distribution behind Redshift `random()` function?

I'm using AWS Redshift's random() function to select a small portion of data from a giant dataset. But feel it's safer to understand how its random function works.

It's uniform distribution or normal distribution or any other method to randomly choose the items?