Only the multi-master, multi-region DynamoDB answer makes sense. Multi-AZ deployments do not provide sufficient availability when a business loses USD

360,000 per hour of unavailability. As RDS does not natively support multi-region, and ACID does not perform well/at all over large distances between regions, only the DynamoDB answer works.

Reference:

http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/Streams.CrossRegionRepl.html

Global secondary index ג€" an index with a partition key and a sort key that can be different from those on the table. A global secondary index is considered "global" because queries on the index can span all of the data in a table, across all partitions.

Reference:

http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/SecondaryIndexes.html

Note that the bootstrapping process can be slower if you have a complex application or multiple applications to install. Managing a fleet of applications with several build tools and dependencies can be a challenging task during rollouts. Furthermore, your deployment service should be designed to do faster rollouts to take advantage of Auto Scaling. Prebaking is a process of embedding a significant portion of your application artifacts within your base AMI. During the deployment process you can customize application installations by using EC2 instance artifacts such as instance tags, instance metadata, and Auto Scaling groups.

Reference:

https://d0.awsstatic.com/whitepapers/overview-of-deployment-options-on-aws.pdf

AWS specifically recommends Blue-Green for super-fast, zero-downtime deploys - and thus rollbacks, which are redeploying old code. You use various strategies to migrate the traffic from your current application stack (blue) to a new version of the application (green). This is a popular technique for deploying applications with zero downtime.

Reference:

https://d0.awsstatic.com/whitepapers/overview-of-deployment-options-onaws.pdf

Linux stacks can install custom cookbooks from any of the following repository types: HTTP or Amazon S3 archives. They can be either public or private, but

Amazon S3 is typically the preferred option for a private archive. Git and Subversion repositories provide source control and the ability to have multiple versions.

Reference:

http://docs.aws.amazon.com/opsworks/latest/userguide/workingcookbook-installingcustomenable.html

Since you are not using VPC endpoints, outbound requests for the code sitting in S3 are routed though the NAT for the VPC's private subnets. If this networking fails, runtime bootstrapping through code download will fail due to network unavailability and lack of access to the Internet, and thus Amazon S3.

Reference:

http://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/VPC_NAT_Instance.html

Oracle is not open source, and requires a bring your own license model.

Reference:

http://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/CHAP_Oracle.html

The ceiling throughput for PIOPS on EBS is 320MiB/s.

Reference:

http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EBSVolumeTypes.html

When a user is trying to detach an EBS volume, the user can either terminate the instance or explicitly remove the volume. It is a recommended practice to unmount the volume first to avoid any file system damage.

A user can always create a new EBS volume of a higher size than the original snapshot size. The user cannot create a volume of a lower size. When the new volume is created the size in the instance will be shown as the original size. The user needs to change the size of the device with resize2fs or other OS specific commands.