A company is migrating latency-sensitive HPC workloads that currently use on-premises NAS file shares. The company needs to provide NFS and SMB multi-protocol access and achieve the LEAST possible latency in AWS. Which combination of actions will meet these requirements with the least latency? (Choose two.)

For latency-sensitive HPC workloads, a cluster placement group ensures that EC2 instances are physically close together within an Availability Zone, minimizing network latency between them. FSx for NetApp ONTAP supports multi-protocol access (NFS and SMB) and offers high-performance, low-latency file storage, making it suitable for demanding HPC workloads. Attaching EC2 instances to an Amazon FSx for Lustre file system is incorrect because while Lustre is high-performance, it primarily supports POSIX-compliant file access (NFSv3, Lustre client) and doesn't natively offer SMB multi-protocol access as a primary feature like FSx for ONTAP. A partition placement group is designed for fault tolerance across racks, not for the lowest possible latency between instances. FSx for OpenZFS is high-performance but is not explicitly designed for the multi-protocol (NFS and SMB) requirements in the same way FSx for ONTAP is.