AWS System Administration

Chapter 12. Backups

If one relies solely on marketing materials provided by cloud hosts and resellers, one might be forgiven for thinking that the cloud is a magical place where nothing breaks and everything Just Works. Unfortunately, this is not the case. Cloud-based infrastructure requires just as much backup planning as a traditional self-hosted architecture—sometimes more, because of the dynamic nature of the cloud. Fortunately there is a silver lining: along with new challenges, the cloud provides new features to make backups simpler and reduce implementation time. For example, ticking a checkbox is figuratively all that it takes to set up scheduled backups for the RDS service, as shown in Figure 12-1.

Figure 12-1. Built-in cloud backup facility in the AWS Relational Database Service

Although business types can think of the cloud as a single logical entity, we must look beyond that high-level presentation to view our cloud as a series of datacenters spread across multiple regions, and plan our backups accordingly. To run a highly available service, you would not put all of your servers in a single datacenter. You should plan backups with the same consideration in mind.

Furthermore, it is also important to think about off-site backups. When working with AWS, this can mean one of two things: storing backups outside your primary region, or going a step further and storing them entirely ...