AWS RDS / Aurora Feature Summary
RDS Aurora up to 64TB of storage per instance / database / cluster
Aurora like other RDS databases uses Database Parameter groups to control configuration
Shared nothing model
Uses in memory buffer data and dictionary caches with CloudWatch based metrics collection including buffer cache hit rate
Cloudwatch provides much more than just cache hit rates for Aurora
One primary and up to 15 read replicas – read replicas need to be explicitly used
Aurora High Availability, Recovery and Cloning
Read replica is promoted to primary in case of primary failure route 53 automatically updates DNS
Can use Aurora Backtrack for point in time recovery of same instance
Can use Aurora Snapshots & Continuous Backups for point in time recovery to another instance – RDS Console shows last point in time recovery date
Aurora DB Cloning Uses Copy on Write Technology
Aurora Scaling, Migration and Pricing
Aurora can use Push button vertical scaling by DB Instance Type – up or down
Migrations can use DMS and SCT
Aurora is more expensive than AWS RDS based MySQL or PostgreSQL solutions but if you need scalability/performance and reliability and you don’t want to build it yourself – then Aurora is a very good choice. It may even scale better than Oracle and in the AWS cloud running RDS is less expensive than Oracle.
Aurora is either Postgres or MySQL compatible – not both simultaneously – you choose at instance / database creation time
More from LonzoDB on AWS
References: