Which solution will complete the migration to AWS in the LEAST amount of time?
Export the on-premises VMs and copy them to an Amazon S3 bucket. Use VM Import/Export to create AMIs from the VM images that are stored in Amazon S3. Order an AWS Snowball Edge device. Copy the NFS server data to the device. Restore the NFS server data to an Amazon EC2 instance that has NFS configured.
Configure AWS Application Migration Service with a connection to the VMware cluster. Create a replication job for the VMS. Create an Amazon Elastic File System (Amazon EFS) file system. Configure AWS DataSync to copy the NFS server data to the EFS file system over the Direct Connect connection.
Recreate the VMs on AWS as Amazon EC2 instances. Install all the required software packages. Create an Amazon FSx for Lustre file system. Configure AWS DataSync to copy the NFS server data to the FSx for Lustre file system over the Direct Connect connection.
Order two AWS Snowball Edge devices. Copy the VMs and the NFS server data to the devices. Run VM Import/Export after the data from the devices is loaded to an Amazon S3 bucket. Create an Amazon Elastic File System (Amazon EFS) file system. Copy the NFS server data from Amazon S3 to the EFS file system.
Explanations:
While using AWS Snowball Edge can help with large data transfers, exporting and copying VMs to S3, then using VM Import/Export, is generally slower than using direct replication methods. This option may also lead to additional downtime during the export/import process, making it less efficient for a large number of VMs.
AWS Application Migration Service allows for efficient and quick migration of VMs through replication, minimizing downtime. Using AWS DataSync over the Direct Connect connection to copy the NFS data to Amazon EFS optimizes the transfer speed, making this the fastest overall solution for both VMs and NFS data.
Recreating VMs manually is time-consuming and prone to errors, especially with custom software installations. Although using FSx for Lustre is efficient for high-performance file storage, the manual setup of instances and software negates the time savings compared to automated replication methods.
Ordering two AWS Snowball Edge devices would add logistical delays and the process of using VM Import/Export after copying the data is slower than using replication methods. Additionally, this option does not leverage the benefits of the Direct Connect connection efficiently for the NFS data transfer.
From what I’ve seen, the answer is:
Configure AWS Application Migration Service with a connection to the VMware cluster. Create a replication job for the VMS. Create an Amazon Elastic File System (Amazon EFS) file system. Configure AWS DataSync to copy the NFS server data to the EFS file system over the Direct Connect connection.