What changes should the company make to meet these requirements?
Establish a second DX connection for redundancy. Use DynamoDB global tables to replicate data to a second Region. Modify the application to fail over to the second Region.
Use an AWS managed VPN as a backup to DX. Create an identical DynamoDB table in a second Region. Modify the application to replicate data to both Regions.
Establish a second DX connection for redundancy. Create an identical DynamoDB table in a second Region. Enable DynamoDB auto scaling to manage throughput capacity. Modify the application to write to the second Region.
Use AWS managed VPN as a backup to DX. Create an identical DynamoDB table in a second Region. Enable DynamoDB streams to capture changes to the table. Use AWS Lambda to replicate changes to the second Region.
Explanations:
Establishing a second DX connection for redundancy ensures that the network connection remains available, and using DynamoDB global tables allows for automatic, multi-region data replication, satisfying the requirement for high availability and fault tolerance. Modifying the application to fail over to the second region provides a complete solution for maintaining data availability.
Using an AWS managed VPN as a backup to DX does not meet the requirement for consistent speeds greater than 2 Gbps, as VPNs typically have lower bandwidth capabilities. Creating an identical DynamoDB table in a second region without automatic replication would not ensure data consistency or high availability across regions.
While establishing a second DX connection provides redundancy, creating an identical DynamoDB table in a second region does not inherently provide fault tolerance or high availability. Writing to the second region would require application changes and may lead to data inconsistencies. Auto scaling alone does not address the replication requirement across regions.
Using an AWS managed VPN as a backup does not guarantee the required bandwidth and consistency needed. Creating an identical DynamoDB table in a second region and using DynamoDB streams with Lambda for replication would involve delays and potential data loss, as the streams capture changes after they occur, which does not ensure immediate consistency across regions.
I estimate that the answer is:
Establish a second DX connection for redundancy. Use DynamoDB global tables to replicate data to a second Region. Modify the application to fail over to the second Region.