Question

Difficulty: MediumHigh-Performing and Scalable Storage Solutions

A company is deploying a machine learning inference application on a fleet of Amazon EC2 instances distributed across three Availability Zones. The instances require concurrent read-write access to a shared repository of model weights and input datasets. The storage solution must be POSIX-compliant, support sub-millisecond read latencies, and scale dynamically to handle peak throughput of up to 6 GB/s6\text{ GB/s}.

Which combination of storage solutions and configurations will meet these requirements? (Select TWO.)

  1. Create an Amazon EFS file system configured with Elastic throughput mode.Answer
  2. Create an Amazon FSx for NetApp ONTAP file system in a Multi-AZ deployment.Answer
  3. C
    Provision an Amazon EBS gp3 volume and enable Multi-Attach to share the block storage across all three Availability Zones.
  4. D
    Provision an Amazon EBS io2 volume with Multi-Attach enabled and attach it to the EC2 instances in different Availability Zones.
  5. E
    Deploy an Amazon EFS file system using the EFS One Zone storage class.

Answer

The correct configurations are creating an Amazon EFS file system configured with Elastic throughput mode and deploying an Amazon FSx for NetApp ONTAP file system in a Multi-AZ configuration.
The correct options are the Amazon EFS file system with Elastic throughput mode and the Amazon FSx for NetApp ONTAP file system in a Multi-AZ deployment. Amazon EFS is a fully managed, shared POSIX-compliant file system that spans multiple Availability Zones. When configured with Elastic throughput mode, EFS automatically scales to meet high-throughput workloads (up to 10 GB/s10\text{ GB/s} read throughput) without manual provisioning, making it ideal for unpredictable workloads. Amazon FSx for NetApp ONTAP is another fully managed, shared storage service that supports POSIX-compliant protocols (NFS and SMB), spans multiple Availability Zones, and offers high-performance throughput of several gigabytes per second with sub-millisecond latencies using NVMe caches.

Step-by-Step Solution

1
Analyze the requirements for shared access, availability, and network reachability.
The application runs on EC2 instances across three Availability Zones and requires a shared, POSIX-compliant file system. This rules out standard single-instance EBS volumes and non-POSIX storage solutions.
This establishes that a multi-AZ shared file storage service like Amazon EFS or Amazon FSx is required to allow concurrent read-write access.
2
Evaluate Amazon EBS Multi-Attach capabilities for multi-AZ applications.
EBS volumes (both gp3 and io2) are single-AZ resources. Even with Multi-Attach enabled on io2, they cannot span across different Availability Zones.
This eliminates the EBS-based options because they cannot meet the multi-AZ network requirement.
3
Evaluate the throughput and availability requirements against Amazon EFS and Amazon FSx configurations.
Amazon EFS with Elastic throughput mode provides automatically scaling throughput up to 10 GB/s10\text{ GB/s} across multiple Availability Zones. Amazon FSx for NetApp ONTAP Multi-AZ provides high-throughput shared storage that spans multiple Availability Zones. EFS One Zone is eliminated because it does not support multi-AZ availability.
This identifies the two correct, scalable, and high-performing storage configurations.

Key Concept

High-performing shared storage solutions spanning multiple Availability Zones
Rate this question