Skip to content

Exploring AWS M5ad Instances: General Purpose with Local NVMe Storage

When it comes to Amazon EC2 instances, AWS offers a wide range of flavors tailored to specific needs. For workloads that require both general-purpose compute power and fast local storage, the M5ad instances stand out. Built on AMD EPYC processors and equipped with NVMe-based SSD storage, they deliver a balance of price, performance, and storage flexibility.

What Are M5ad Instances?

The M5ad family belongs to AWS’s M5 general-purpose line, similar to M5a instances, but with an added twist:

  • They are powered by AMD EPYC 7000 series processors (up to 2.5 GHz).
  • They include local NVMe SSD instance storage, physically attached to the host server.

This makes them a great option for workloads that need low-latency, high-speed storage in addition to balanced compute and memory resources.

Key Features

  • Processor: AMD EPYC 7000 series (up to 2.5 GHz)
  • vCPUs: 2 – 96, depending on the instance size
  • Memory: 8 GiB per vCPU (same ratio as M5a)
  • Local Storage: NVMe-based SSDs included with each instance
  • EBS-Optimized: Support for high-throughput Amazon Elastic Block Store (EBS)
  • Networking: Up to 25 Gbps with the Elastic Network Adapter (ENA)
  • Built on Nitro: AWS Nitro System for improved security and performance

Instance Sizes

Like M5a, M5ad instances come in various sizes, but each size also includes local NVMe SSD storage:

  • m5ad.large – 2 vCPUs, 8 GiB RAM, 1 × 75 GB NVMe SSD
  • m5ad.xlarge – 4 vCPUs, 16 GiB RAM, 1 × 150 GB NVMe SSD
  • m5ad.2xlarge – 8 vCPUs, 32 GiB RAM, 1 × 300 GB NVMe SSD
  • m5ad.4xlarge – 16 vCPUs, 64 GiB RAM, 2 × 300 GB NVMe SSDs
  • m5ad.8xlarge – 32 vCPUs, 128 GiB RAM, 2 × 600 GB NVMe SSDs
  • m5ad.12xlarge – 48 vCPUs, 192 GiB RAM, 2 × 900 GB NVMe SSDs
  • m5ad.16xlarge – 64 vCPUs, 256 GiB RAM, 4 × 600 GB NVMe SSDs
  • m5ad.24xlarge – 96 vCPUs, 384 GiB RAM, 4 × 900 GB NVMe SSDs

Use Cases

M5ad instances are best suited for applications that require both compute and low-latency storage. Common scenarios include:

  • Caching and temporary data storage – Fast NVMe storage is ideal for short-lived or temporary data.
  • Data logging and processing – Local SSDs help handle bursty I/O workloads.
  • Gaming servers – Benefit from high-speed storage and compute balance.
  • Media processing and rendering – NVMe storage helps with large files and temporary data.
  • Databases – Useful for workloads where fast local storage is beneficial, such as NoSQL or cache databases.

Benefits of Choosing M5ad

  1. Cost Savings – Like M5a, M5ad is typically up to 10% cheaper than Intel-based M5d instances.
  2. Local NVMe Storage – High-speed SSD storage for workloads that need fast disk I/O.
  3. Balanced Performance – Equal memory-to-vCPU ratio as M5/M5a, making them versatile.
  4. Scalability – Available in sizes suitable for both small and large workloads.

M5ad vs. Other M5 Variants

  • M5a – Cheaper general-purpose instances, but no local NVMe storage.
  • M5ad – Same as M5a, but with NVMe SSDs included.
  • M5d – Intel-based equivalent with NVMe storage, generally more expensive.

Things to Keep in Mind

  • Ephemeral Storage: NVMe instance storage is not persistent. Data is lost if the instance stops, hibernates, or terminates. Always use Amazon EBS or S3 for permanent storage needs.
  • Workload Fit: M5ad makes sense only if your workload can benefit from fast local disk access. If not, M5a might be the more cost-effective choice.

Conclusion

AWS M5ad instances combine the affordability of AMD EPYC processors with the performance benefits of local NVMe SSD storage. They provide a strong option for workloads that require both compute flexibility and low-latency storage, such as gaming, caching, and media processing.

If you’re already considering M5a instances but need high-speed local storage, M5ad offers the perfect balance between performance, storage, and cost savings.