Which of the following options can the Solutions Architect design to create a comprehensive solution for this customer that meets the disaster recovery requirements?
Back up the application and database data frequently and copy them to Amazon S3. Replicate the backups using S3 cross-region replication, and use AWS CloudFormation to instantiate infrastructure for disaster recovery and restore data from Amazon S3.
Employ a pilot light environment in which the primary database is configured with mirroring to build a standby database on m4.large in the alternate region. Use AWS CloudFormation to instantiate the web servers, application servers and load balancers in case of a disaster to bring the application up in the alternate region. Vertically resize the database to meet the full production demands, and use Amazon Route 53 to switch traffic to the alternate region.
Use a scaled-down version of the fully functional production environment in the alternate region that includes one instance of the web server, one instance of the application server, and a replicated instance of the database server in standby mode. Place the web and the application tiers in an Auto Scaling behind a load balancer, which can automatically scale when the load arrives to the application. Use Amazon Route 53 to switch traffic to the alternate region.
Employ a multi-region solution with fully functional web, application, and database tiers in both regions with equivalent capacity. Activate the primary database in one region only and the standby database in the other region. Use Amazon Route 53 to automatically switch traffic from one region to another using health check routing policies.
Explanations:
While copying backups to Amazon S3 and using CloudFormation to restore data can be part of a disaster recovery solution, this approach does not meet the RTO of 5 minutes and RPO of 1 hour. The process of restoring backups from S3 and provisioning infrastructure with CloudFormation will likely exceed the required RTO.
The pilot light approach involves maintaining minimal infrastructure in the disaster recovery region. However, the approach of using an m4.large instance for the database and resizing it during a disaster violates the requirement for a 100 GB memory database and may not meet the 5-minute RTO.
This solution meets the disaster recovery requirements with a scaled-down version of the production environment in the alternate region. The database is in standby mode, and Auto Scaling for the web and application tiers allows the environment to scale up quickly during a disaster. This satisfies both the 5-minute RTO and 1-hour RPO.
A multi-region solution with fully functional web, application, and database tiers in both regions is too costly and complex for this scenario. It also does not allow for the 100 GB database to be efficiently handled with the given memory requirements, making it not a cost-effective solution.