Azure Virtual Machines Sizes

Topics Covered

Overview

When you use Azure Virtual Machines (VMs), their sizes matter. An Azure Virtual Machines Sizes decides how much power it has for things like memory, processing, and storage. Azure offers different Azure Virtual Machines Sizes, each with its own abilities. Some VMs are like small laptops, while others are like powerful servers. So, while choosing an Azure VM, remember to select the appropriate size according to your need and budget. It's like choosing the right shirt size as you want it to fit perfectly.

What is an Azure VM?

Azure Virtual Machine enables us to create and use virtual computers in the cloud, like a service, without worrying about the physical stuff. It possesses a CPU, memory, storage space for your files, and the ability to access the internet if required. Virtual machines (VMs) are generally thought of as software-defined computers or virtual computers within physical servers, existing entirely as code.

Azure Virtual Machines allows you to install a wide range of computing solutions, from simple development and testing environments to complex, production-ready applications. When you create an Azure Virtual Machine, you generally specify parameters such as the preferred operating system, VM size, storage options, networking setup, and more. Once created, you can access and control the VM remotely, just like a physical computer. Here are a few ways in which virtual machines can be used:

  • Developing and deploying applications on the cloud.
  • Experimenting with new operating systems, even beta versions.
  • Making a copy of your current operating system for backup.
  • Using an older operating system to handle virus-infected data or run outdated software.
  • Creating a fast and easy setup for developers to test their code.

Azure VM Types

In Azure, Microsoft offers a variety of virtual machine sizes organized into groups known as "series." Each series is designed to serve a specific purpose or use case.

azure vm types

Azure divides virtual machines into five distinct groups, outlined in the table below.

Type of VMSizesDescription
General purposeB, Dsv3, Dv3, Dasv4, Dav4, DSv2, Dv2, Av2, DC, DCv2, Dpdsv5, Dpldsv5, Dpsv5, Dplsv5, Dv4, Dsv4, Ddv4, Ddsv4, Dv5, Dsv5, Ddv5, Ddsv5, Dasv5, Dadsv5These are the well-balanced ratio of CPU to memory. It is suitable for tasks like testing, development, managing small to medium databases, and hosting web servers with moderate traffic.
Compute optimizedF, Fs, Fsv2, FXThese VMs provide a high CPU-to-memory ratio. Well-suited for tasks such as hosting medium-traffic web servers, handling batch processes, managing network appliances, and supporting application servers.
Memory optimizedEsv3, Ev3, Easv4, Eav4, Epdsv5, Epsv5, Ev4, Esv4, Edv4, Edsv4, Ev5, Esv5, Edv5, Edsv5, Easv5, Eadsv5, Mv2, M, DSv2, Dv2These VMs provide a high memory-to-CPU ratio. It is ideal for roles like hosting relational database servers, managing medium to large caches, and performing in-memory analytics.
Storage optimizedLsv2, Lsv3, Lasv3These VMs provide high disk throughput and IO, perfect for tasks involving Big Data, SQL and NoSQL databases, data warehousing, and managing large transactional databases.
GPUNC, NCv2, NCv3, NCasT4_v3, NC A100 v4, ND, NDv2, NGads V620, NV, NVv3, NVv4, NDasrA100_v4, NDm_A100_v4These VMs are designed for specific tasks like demanding graphic rendering, deep learning model training, video editing and inferencing (ND) that require high performance. These specialized virtual machines have single or multiple GPUs to handle the workload efficiently.
High performance computeHB, HBv2, HBv3, HBv4, HC, HXThis is the most powerful CPU virtual machines, with optional high-throughput network interfaces (RDMA).

Note:

  • The first capital letter in the VM name represents the family. For example, in D4ads_v5 VM, the letter D represents that this VM belongs to the D-Series.
  • The first number in name represents the number of vCPU (virtual central processing unit) the VM has. For example, in D4ads_v5 VM, the number 4 represents that it has four vCPUs.
  • Lowercase letters in the VM names denote the additional features.

Azure VM Pricing Models

Let's explore the various VM Pricing Models in detail:

Azure Free Tier

Similar to other cloud platforms, Azure provides a trial with free credits. Azure provides a free tier of $200 in Azure credits for the first 30 days, followed by a capped number of additional free services for the next 12 months.

You have the flexibility to create services in any region that Azure supports. Additionally, you can deploy multiple instances as long as their total number remains within the specified limits.

Pay As You Go

Azure VMs feature a payment approach in which you are charged for every second a VM resource is active. This ensures that you are only charged for the time you utilize.

This is the most versatile option and is best suited for tasks that require continuous operation or short-term workloads. There are no up-front fees or long-term payments required. You may scale your computing capacity up or down based on the needs of your application, and you only pay for the instances you use every hour. However, it is important to mention that prices under this model are often higher than other pricing models.

Reserved VM Instances

When you anticipate using a virtual machine (VM) for more than a year, you can reserve a VM instance, which can result in cost savings of up to 72 percent. Reserving a VM involves making a commitment to a specific location for one or three years.

You have the flexibility to modify your reservations by returning or exchanging them. Additionally, cancelling a reserved instance with an early cancellation fee is possible, provided it doesn't exceed the yearly cap.

Typical scenarios where Reserved Instances (RIs) are valuable include:

  1. Steady Application Usage:
    RIs are beneficial when your application maintains a consistent usage level over time.
  2. Reserved Capacity Needs:
    If your application has a predictable need for computing capacity, RIs can help you secure the required resources.
  3. Cost-Controlled VM Usage:
    Opting for Azure VMs under a one or three-year commitment allows you to lock in and control computing expenses, resulting in cost savings.

With the reserved VM instances framework, you have the following flexibility:

  1. Modify Reservations:
    Adapt and adjust reservations to accommodate your expanding needs by cancelling or exchanging them.
  2. Plan and Budget:
    With upfront payments for one or three years, you can improve your financial planning and enable better cost forecasting.
  3. Priority Capacity:
    Enjoy prioritized access to compute capacity within Azure regions, ensuring your resources are readily available.
  4. Flexible Instance Management:
    Benefit from automatic management of Azure RIs by customizing your configuration to the demands of the workload.
  5. Monthly Payment Option:
    Embrace a monthly payment approach for reserved VMs without any added costs, avoiding the need for upfront payments.

Spot Virtual Machines

Azure VM spot pricing is an excellent alternative if you have a tight budget, don't necessarily need a hundred percent uptime, or have the automation to deal with VM availability concerns.

Azure data centres always have some unused capacity. To ensure data centres are being efficiently used, Azure created spot pricing, which grants discounts of up to ninety percent on pay-as-you-go VM pricing. There are no upfront or long-term payment commitments.

Azure retains the authority to terminate or power down spot VMs whenever the data center necessitates additional computing resources or if the ongoing price surpasses the initially agreed-upon rate. This circumstance poses challenges for running production workloads on spot VMs.

Spot Virtual Machines has the following applications according to various types of workloads:

  1. Spot Virtual Machines can be used in development and testing environments, including continuous integration and continuous delivery workloads.
  2. It is used in tasks involving batch processing, visual rendering applications, and specific high-performance computing needs.
  3. It can also be used in workloads such as analytics, big data processing, stateless applications at a large scale, and applications based on containers.

Azure Virtual Machine Pricing Examples

Let's take a look at the pricing options for several types of Linux virtual machines in the West US region.

VMHardwarePay as you go priceSpot priceThree year reserved instance price
D2s v3 VM2 vCPUs, 8 GB RAM and 16 GB local storage$0.09600/hour$0.0497/hour$0.0369/hour
D8s v3 VM8 vCPUs, 64 GB RAM and 128 GB local storage$0.384/hour$0.1986/hour$0.1474/hour
D32s v3 VM32 vCPUs, 128 GB RAM and 256 GB local storage$1.536/hour$0.7942/hour$0.5896/hour
D64s v3 VM64 vCPUs, 256 GB RAM and 512 GB local storage$3.072/hour$1.5883/hour$1.1792/hour

Conclusion

  • A virtual machines (VMs) are a fundamental component of cloud computing. They enable the creation of a cloud-based computer instance that may be used to host applications and process data.
  • Azure Virtual Machines allows you to install a wide range of computing solutions, from simple development and testing environments to complex, production-ready applications.
  • When you use Azure Virtual Machines (VMs), Azure Virtual Machines Sizes sizes matter. An Azure Virtual Machines Sizes decides how much power it has for things like memory, processing, and storage.
  • Similar to other cloud platforms, Azure provides a trial with free credits.
  • Azure VMs feature a payment pay-as-you-go approach in which you are charged for every second a VM resource is active.
  • When you anticipate using a virtual machine for more than a year, you can reserve a VM instance.
  • Azure VM spot pricing is an excellent alternative if you have a tight budget, don't necessarily need a hundred percent uptime.