Disaster Recovery

Cloud computing really sped up a lot the adoption of disaster recovery strategies for companies of all sizes. Specially because of the cloud low costs and the agility to test the DR, SMB companies started to implement disaster recovery policies.

Here are some recommend strategies for DR.

AWS Amazon Cloud

Disaster Recovery Backup Off-Site

In this model, all the company data is stored as a cloud backup. If you have a problem on your production site, the cloud backup data will be used to create new servers (mainly using server images). This model offers the lowest cost (just backup) and the highest recovery time.

Disaster Recovery Backup Off-Site

AWS Amazon Cloud

Disaster Recovery Pilot Light

The strategy here consists on maintaining a running server as a read only database replica in the cloud, and the other servers turned off. In case of problems in the production site, the servers are started and the database server has its size increased becoming a read write database to support the application traffic.

Disaster Recovery Pilot Light

AWS Amazon Cloud

Disaster Recovery Warm StandBy

In this model, all the servers are replicated to the cloud, they stay on there anytime, but in a lower size (the minimum to work as a stand by). In case of failure in the production site, the servers are scaled up to receive the application traffic.

Disaster Recovery Warm StandBy

AWS Amazon Cloud

Disaster Recovery Multi-Site

Here, the strategy in to replicate everything to the cloud, all the servers stay up and receiving production traffic. In this way, both environments (cloud and on premise) are in production, and in case of a failure of any of them, the other takes care of the whole workload. In this scenario you have the highest cost, and the lowest recovery time (almost zero).

Disaster Recovery Multi-Site

AWS Amazon Cloud

Agent Replication

In this model, we utilize third part tools that replicate the servers’ data to the cloud using block level replication. Those tools use agent installed inside the on premise servers to replicate the data to the cloud in almost real time. This solution also orchestrates the test process for disaster recovery, making everything easier.
As this scenario uses a transparent replication of the servers, the DR strategy is sometimes faster and more assertive, and you don’t need to have the servers running in the cloud.

Agent Replication