Changes to connection handling with Aurora Postgres?

Were any changes made to how connections are handled between PostgreSQL and Aurora Postgres? Specifically, there are limitations in Postgres with connections due to memory usage, and a customer wants to know if any improvements were made with Aurora.

In addition, if the connection "math" is the same as vanilla Postgres, is the total max connections for a cluster equal to the # of replicas * max connections per, or something else?

Thanks for any insights!