New Training: Planning for Amazon RDS Failover

In this 6-video skill, CBT Nuggets trainer Bart Castle discusses failover and high availability options for Amazon RDS. Gain an understanding of how to design networks that support RDS Multi-AZ deployments as well as how to create and work with Read Replicas. Watch this new AWS training.

Watch the full course: Amazon Relational Database Service

This training includes:

• 6 videos

• 57 minutes of training

You’ll learn these topics in this skill:

• Plan for Amazon RDS Failover

• Amazon RDS Failover Patterns

• Deploy Amazon RDS Multi-AZ

• Force Multi-AZ Failover

• Create RDS Read Replica

• Promote a RDS Read Replica

How Read Replicas Removes Read/Write Bottlenecks on AWS Databases

One of the primary concerns with scaling an environment is performance bottlenecks. As organizations grow and more users begin requesting information from services using back-end databases to store information, bottleneck issues can arise as the database struggles to support the volume of read requests. This is where creating Read Replica databases can help in redistributing the database read requests in such a way that removes these bottleneck challenges.

Read Replicas are one-to-one image-level copies of the original Amazon RDS database designed in a read-only format so that data can be read from these databases. By reconfiguring the environment, read requests can be evenly distributed to the original database and the Read Replica databases to alleviate strain from the original database.

As updates are applied to the production environment, the original database will get updated with new database entries and the replicas will get updated through the replication process so they can facilitate up-to-date read requests passed off by the original database.

It should be noted that read replicas have other uses, too. For instance, they can be used in a disaster recovery scenario. Let's consider AWS Multi-AZ deployments for a moment. Multi-AZ deployments provide advanced availability for databases within the same region. Should there be an infrastructure failure, RDS can automatically failover to the standby instance while minimizing downtime.

Think of using Multi-AZ as a sort of RAID 1 array for your databases. Though RAID should never be considered a backup solution, it can help mitigate disasters and provide higher availability by providing mirrored replicas that your environment can failover to.