We talked about performances issues with uuid and cuid last month. This time we will dig deeper into the mess of uuid, and because sometimes integer ids are not recommended, here is an introduction to nanoid instead. Id like to have a benchmark on it too, but seems for now that its way more efficient.

Should we use integer, cuid or uuid for our database ids type ? An interesting comparison between the 3 of them based on their performances, and by extension, their costs.