Aurora is proprietary and only available on AWS. Any drivers you use for MySQL and PostgreSQL will work for it. It has 5x the performance of MySQL and and 3x the performance of PostgreSQL. Aurora supports up to 15 read replicas and 128 TB of storage, the storage and read replicas scale with size with storage increments down to 10 GB, you can even have a policy in place similar to RDS DB scaling for your read replicas, such as creating a new read replica if CPU usage is above 60% on your existing RRs. Aurora ensures HA by having 6 copies of your data across 3 AZ, only 4 out of 6 are need to be able to write and 3 out of 6 to be able to read. Similar to RDS you have one “master instance” that writes and any of the read replicas can be promoted to the master instance. The master instance is responsible for writing and has a writer endpoint DNS name that can be accessed. The read replicas are under a reader endpoint DNS name that essentially acts as a load balancer between all of the read replicas, as the read replicas automatically scale they are added to this LB. You have advanced monitoring, zero-downtime patching, and a feature called backtrack that allows you to restore your DB to any point in time without the use of snapshots. Aurora has sub 30 second failover when your write replica fails and sub 10 ms replication lag between the master instance and the read replica.