Designing High-Performing Database Solutions

1. What is the advantage of using RDS to set up a database?
The advantage of using RDS to set up a database is that after you describe the infrastructure and your needs to AWS, the database infrastructure is automatically set up and maintained for you; even failover is automatically managed. All you have to do is work with your data records.

2. What is the disadvantage of using RDS to set up a database?
The disadvantage of using RDS to set up most database types is that the standard deployment of RDS is a primary and standby database design. Another disadvantage is that RDS supports a set number of database engines; there’s no flexibility. There could also be a disadvantage if you are looking for more of an active-active deployment. Of course, nothing stops you from building any database infrastructure design that you wish by using EC2 instances. Or perhaps you could look at Amazon Aurora, which is officially part of RDS but really something quite different (see Question 6).

3. How can read replicas help improve database performance?
Read replicas can help improve database performance by taking the load off the primary database instance in regard to queries. A typical RDS deployment has a primary database and a standby database; however, the standby database stands by and waits for disaster and does nothing except make sure that it’s up to date with the primary database instance. That’s all well and good, but perhaps your primary database is becoming bogged down by handling all the queries by itself. Adding read replicas in regions where they are needed can increase performance as the read replicas can handle the queries in those specific regions.

4. What two options are available at AWS for hosted databases with global multi-region solutions?
AWS has two database solutions for multi-region deployments. The first is DynamoDB, a NoSQL solution that supports global tables that can span multiple AWS regions. AWS also has an SQL solution, Amazon Aurora, which can also span multiple AWS regions with a global deployment. The only other option is a database solution that you build yourself using EC2 instances.

5. What is the difference between eventual consistency and strong consistency?
The difference between eventual and strong consistency in regard to database storage—specifically with DynamoDB—is that you have a choice between living with the reality of data replicated to multiple locations (eventual consistency) and having checks made to all the storage partitions to see which one has the most current copies of data so that those copies can be presented to the application (strong consistency).

6. How does Aurora have an advantage over a standard MySQL deployment with RDS?
Amazon Aurora has a huge advantage over a standard MySQL deployment because the data storage architecture uses a virtual SAN composed of SSDs. The data is also stored in a cluster with multiple writers. As a result, the performance cannot really be compared; Aurora is much faster.

7. Where are continuous backups stored for all AWS database servers?
Continuous backups for all database solutions are stored in AWS as S3 storage.

8. What is an advantage of using ElastiCache to store user state?
The advantages of using ElastiCache to store user state information are speed—after all, the cache storage is operating in RAM—and reliability. Rather than storing the user state directly on the EC2 instance where the user session is taking place, the user’s session information is stored in another location just in case the EC2 instance that the end user is communicating with fails. If there is a failure, and the user begins communicating with another web server, the user’s session information can be retrieved from ElastiCache, allowing the continuation of the user’s session on the other server.