I'm studying Glue Data Catalog and I'm a bit confused. I know that Glue Data Catalog doesn't store the data itself but just the metadata for the databases and tables, but still, where is the location in which they are stored?
Based on publicly available information, Glue Catalog is a drop in replacement to Hive Metatore. It is highly likely that the architecture is very similar to Hive Metastore. At a high level, the Hive metastore consists of a Java server and a MySQL database. The admin manual provides more details.
So the metadata must be stored in a MySQL database managed internally by AWS. This is speculation as there is no concrete information available publicly.