Disaster recovery on AWS: strategies for SMBs

Protect your business against disruptions

A disaster recovery plan defines how your business resumes activity after a major outage: hardware failure, human error, cyberattack or natural disaster. On AWS, several strategies exist with different levels of protection and cost. The choice depends on two metrics: the RTO (acceptable recovery time) and the RPO (acceptable maximum data loss).

The four disaster recovery strategies

Backup and restore

The simplest and least expensive strategy. Your data is backed up regularly to Amazon S3 in another region. In case of a disaster, you recreate the infrastructure and restore the data. The RTO is several hours (the time to recreate the infrastructure) and the RPO depends on the frequency of your backups.

This strategy suits non-critical applications and SMBs with a limited budget. The cost is limited to the S3 storage of the backups, a few euros per month for most SMBs.

Pilot light

A minimal infrastructure runs permanently in a second AWS region: a replicated database, network configuration ready. In case of a disaster, you start the application servers and switch the traffic over. The RTO is 30 minutes to 2 hours and the RPO is a few minutes thanks to the continuous replication of the database.

The cost is moderate because only the database and a few network components run permanently. The application servers are started only when needed.

Warm standby

A reduced version of your infrastructure runs permanently in the second region. The application servers are active but with reduced capacity. In case of a disaster, you increase the capacity and switch the traffic over. The RTO is a few minutes and the RPO is nearly zero.

Active-active (multi-site)

Your infrastructure runs at full capacity in two regions simultaneously. The traffic is distributed between the two regions. In case of a region failure, the other absorbs all the traffic automatically. The RTO and the RPO are nearly zero, but the cost is double your normal infrastructure.

AWS services for disaster recovery

AWS Backup centralizes backup management for EC2, RDS, DynamoDB, EFS and S3. You define backup policies (frequency, retention, cross-region replication) and AWS Backup runs them automatically.

Amazon RDS supports multi-AZ replication (high availability within a region) and cross-region read replicas (disaster recovery). Aurora offers global replication with automatic failover in less than a minute.

Route 53 manages DNS failover with health checks. If your main infrastructure stops responding, Route 53 automatically redirects the traffic to the backup infrastructure.

Test your plan regularly

An untested disaster recovery plan is a plan that does not work. Schedule quarterly tests that simulate a disaster and measure the real RTO and RPO. Document the results and adjust the plan accordingly.

LCMH designs disaster recovery plans on AWS tailored to the budget and requirements of SMBs in Alsace.

For cost optimization, read our article on the 7 AWS cost optimization levers.

Sources

1. AWS, Disaster Recovery of Workloads on AWS. docs.aws.amazon.com/whitepapers/latest/disaster-recovery-workloads-on-aws
2. AWS, AWS Backup. aws.amazon.com/backup
3. AWS, Reliability Pillar - Well-Architected Framework. docs.aws.amazon.com/wellarchitected/latest/reliability-pillar