What is the MOST cost-effective backup solution that will meet all requirements?
Back up all the data to a large Amazon EBS volume attached to the backup media server in the production region. Run automated scripts to snapshot these volumes nightly, and copy these snapshots to the disaster recovery region.
Back up all the data to Amazon S3 in the disaster recovery region. Use a lifecycle policy to move this data to Amazon Glacier in the production region immediately. Only the data is replicated; remove the data from the S3 bucket in the disaster recovery region.
Back up all the data to Amazon Glacier in the production region. Set up cross-region replication of this data to Amazon Glacier in the disaster recovery region. Set up a lifecycle policy to delete any data older than 60 days.
Back up all the data to Amazon S3 in the production region. Set up cross-region replication of this S3 bucket to another region and set up a lifecycle policy in the second region to immediately move this data to Amazon Glacier.
Explanations:
This option only backs up data to Amazon EBS in the production region and does not provide a durable copy in a different geographic region for disaster recovery. While snapshots can be created, they may not meet the immediate restoration requirement at the production site. Additionally, using EBS is generally more expensive than S3 or Glacier for long-term storage.
This option proposes backing up data to Amazon S3 in the disaster recovery region and immediately moving it to Amazon Glacier in the production region. However, it fails to provide immediate access to backups in the production region as the data is stored in Glacier, which has retrieval times that do not meet the 24-hour requirement for disaster recovery access. Moreover, removing data from S3 in the disaster recovery region would not satisfy the requirement of storing data in two locations.
This option suggests backing up to Amazon Glacier in the production region and replicating to Glacier in the disaster recovery region. While it provides cross-region redundancy, Glacier is primarily intended for long-term storage with slower retrieval times, which does not meet the requirement for immediate restoration at the production site. Furthermore, the restoration process from Glacier can take several hours, making it unsuitable for the need for immediate availability.
This option backs up data to Amazon S3 in the production region, which provides immediate access. It then sets up cross-region replication to another S3 bucket in a different region, ensuring the backups are stored durably in two locations. The lifecycle policy to move the data to Amazon Glacier in the second region allows for cost-effective long-term storage while still meeting the requirement of having backups available within 24 hours at the disaster recovery location.
I arrange that the answer is:
Back up all the data to Amazon S3 in the production region. Set up cross-region replication of this S3 bucket to another region and set up a lifecycle policy in the second region to immediately move this data to Amazon Glacier.