Which migration strategy meets these requirements?
Create an AWS Database Migration Service (AWS DMS) replication instance. Authorize the public IP address of the replication instance to reach the data warehouse through the corporate firewall. Create a migration task to run at the beginning of the fata freeze period.
Install the AWS SCT extraction agents on the on-premises servers. Define the extract, upload, and copy tasks to send the data to an Amazon S3 bucket. Copy the data into the Amazon Redshift cluster. Run the tasks at the beginning of the data freeze period.
Install the AWS SCT extraction agents on the on-premises servers. Create a Site-to-Site VPN connection. Create an AWS Database Migration Service (AWS DMS) replication instance that is the appropriate size. Authorize the IP address of the replication instance to be able to access the on-premises data warehouse through the VPN connection.
Create a job in AWS Snowball Edge to import data into Amazon S3. Install AWS SCT extraction agents on the on-premises servers. Define the local and AWS Database Migration Service (AWS DMS) tasks to send the data to the Snowball Edge device. When the Snowball Edge device is returned to AWS and the data is available in Amazon S3, run the AWS DMS subtask to copy the data to Amazon Redshift.
Explanations:
Although creating an AWS DMS replication instance and setting it to migrate data during the freeze period might seem viable, the 50 Mbps internet connection would be inadequate to transfer 30 TB of data within a two-week window, making it impractical.
Using AWS SCT extraction agents to extract data and upload it to S3 is a valid approach; however, it doesn’t address the time constraint effectively, as the internet connection speed would still limit the data transfer rate, and it may not complete within the data freeze period.
This option suggests using AWS SCT extraction agents and a Site-to-Site VPN. However, the same limitation applies as in option A: the 50 Mbps internet connection will not suffice for transferring 30 TB within the required timeframe.
This option involves using AWS Snowball Edge, which allows for large-scale data transfer without relying on bandwidth limitations. The data can be extracted and sent to the Snowball device, which is then returned to AWS, where the data is uploaded to S3 and subsequently moved to Redshift. This method is well-suited for the given requirements and constraints.
If I’m correct, the answer is:
Create a job in AWS Snowball Edge to import data into Amazon S3. Install AWS SCT extraction agents on the on-premises servers. Define the local and AWS Database Migration Service (AWS DMS) tasks to send the data to the Snowball Edge device. When the Snowball Edge device is returned to AWS and the data is available in Amazon S3, run the AWS DMS subtask to copy the data to Amazon Redshift.