Azure Kubernetes Service

Just Enough Azure for AKS

Azure Compute Fundamentals

Azure offers a rich portfolio of virtual machine (VM) types and sizes to support workloads from development and testing to mission-critical production systems and large-scale deployments. Most VM series are also compatible with Azure Kubernetes Service (AKS), giving you flexibility in containerized environments.

The image is a diagram titled "Azure Compute" showing different types of virtual machine workloads, including development, test, burstable, small scale systems, mission-critical production, and large systems.

Each VM series is optimized for specific workload profiles—compute-intensive, memory-optimized, storage-optimized, GPU-accelerated, and more. If you’re new to Azure, it’s important to understand these series and their naming conventions. For a complete list of available VM sizes and their specifications, see the official Azure VM sizes reference.

Note

For detailed specifications and regional availability, visit Azure VM Sizes.

The image illustrates different types of Azure Compute virtual machines: compute intensive, memory optimized, storage optimized, and GPU accelerated, each represented by an icon.

Common Azure VM Series and Use Cases

VM SeriesUse CaseExample Workloads
Memory-optimizedHigh memory-to-core ratio, premium disk supportSQL Server, SAP HANA, in-memory caching
General-purposeBalanced CPU-to-memory ratioWeb servers, small to medium databases
Storage-optimizedHigh disk throughput and low-latency I/O with local NVMeBig data, ETL, data warehousing
GPU-acceleratedGraphics-intensive and AI/deep learning tasksModel training, rendering, HPC simulations
ARM-based (Ampere)Cost-effective performance for Linux containersMicroservices, CI/CD agents, edge computing

Choosing the Right VM Series

When selecting a VM series, align its capabilities with your workload requirements:

  • Memory-optimized (E-series)
    Optimal for mission-critical databases and in-memory analytics.
  • General-purpose (D-series)
    Ideal for balanced workloads like web servers and small databases.
  • Storage-optimized (Ls-series)
    Best for applications needing high disk I/O throughput.
  • GPU-accelerated (N-series)
    Designed for AI, machine learning, and graphics rendering.
  • ARM-based (Ampere Altra)
    Provides a favorable performance-to-cost ratio for Linux container workloads.

Understanding Azure VM Naming Conventions

Azure VM names encode key details about the series, size, features, and generation. Decoding these names simplifies the selection process:

Standard_E64ds_v5
│        │ │  └─ v5: Latest version
│        │ └──── d: Premium disk & cache support
│        └──────── 64 vCPUs
└────────────────── Series E: Memory-optimized
  • Standard: Indicates billing tier
  • Series (e.g., E, D, Ls, N)
  • vCPU count
  • Features (d = premium disks, s = support for accelerated networking)
  • Version (v<number>)

The image illustrates Azure Compute with connections to Intel, Ampere, and AMD, featuring the Linux mascot.

Warning

Selecting an unsupported VM size for your AKS cluster can lead to scheduling failures. Always verify compatibility in the AKS VM compatibility guide.


Watch Video

Watch video content

Previous
Azure Landing Zone