A Cloud Cost Optimization Look at Azure Compute

Here’s an overview of cloud operational costs about Azure Compute, specifically with general virtualized machines. We will look at the list pricing of VMs and review a few cloud cost optimization perspectives if your IT teams plan on using Azure Compute at massive scale.

First, what is Azure Compute?

Microsoft Azure features Azure Compute, its core virtualization offering in the cloud. Its focus is to make it easy for developers and engineers to provision and work on new applications or deploy existing ones. Azure Compute provides the infrastructure you need to spin up VMs to run your apps, taking into huge consideration capacity in the cloud and scale on-demand.

Containerization is also available, via tools like Azure Kubernetes Service (AKS), for your applications if that’s part of your cloud infrastructure planning and strategy. Focusing on Windows and Linux VMs, Azure Compute opens up flexible options for migrating VMs to the cloud, which can support enterprise-scale digital transformations.

How much does it cost to run VMs on Azure Compute?

For this exercise, we’ll take a look at the Azure Compute general purpose VM, the D2 v3. Featuring 2 vCPUs and 8GBs of RAM, its pricing is similar to that of AWS’s general purpose M5 EC2 instance. Here’s a look at the cost of running a D2 v3 for one month, running Windows, on the Standard Tier, with 50GB temporary storage, in the West US, East US, and North Central US regions, using the Azure Compute cost calculator.

A look at the D2 v3 VM cost by region

Note: Pricing is as of January 2020, subject to change. Also, pricing here is a cumulative of one month of per-second billing.

With just a regional change alone, the price for the VM changes. This is but one of many variables that can alter the price of cloud computing per month. Other parameters that change this monthly cost include:

• Operating system: It costs more to license Windows than Linux, though you can migrate Windows licenses that your team already owns
• Service tier: Azure offers three tiers for VMs: Standard (for production, higher cost), Basic (for development and testing, lower data transfer, lower costs), and Low Priority, which can alter pricing significantly
• Additional storage: By default, VMs have ephemeral, temporary storage and include one HDD, so adding SSD volumes to VMs adds more cost
• Databases: Azure offers a handful of database services that can be added to VMs, which adds more costs to track
• Support: All VMs include basic service (called Included), and can be up-leveled to higher tiers of support at additional cost

Adding additional features increases the costs of running a VM. While the VMs are billed per second, things like adding SSDs or services have different pricing methods and requirements. SSDs charge by transaction units, for example.

When are you actually accruing costs?

When a VM is running, you will be charged every second. Issuing a stop command to a VM stops its services and the meter. However, additional services, such as storage operations, that might run outside of what the VM needs—e.g. logging, diagnostics—will still incur data transfer charges.

Here’s another quick look at how two D2 v3 VMs can be quite different cost-wise, just by adding features:

Convenient, but complicated

It can be very affordable to spin up a dev/test environment on Azure Compute, but once your teams start putting applications to real-world use in production on multiple VMs within multiple regions, the costs can start to add up. The exercise above only looked at one of a myriad of VM choices. Think of the possibilities!

On Reserved VM Instance pricing

Azure Compute offers a few ways to save on VM costs with their Reserved VM Instances. By paying up-front in either a one-year or three-year term, your teams can save significantly. The draw here is that you need to commit to a much longer timeframe of utilization (e.g. committing to a year is a whole lot different than working with per-second billing!). However, the savings are significant, with 1-year reserved yielding savings of around 18% for the D2 v3 and 3-year yielding about 32% of savings.

Reserved VM Instances are only available for VMs run in Standard mode only. Their prices also change depending on the type of VM used. Knowing all of these complications, it makes sense to speak with your finance or FinOps team to strategically look at actual Azure cloud costs and utilization before making these up-front investments.

Huge savings for teams migrating infrastructure using Windows licenses

Your IT teams can utilize the Azure Hybrid Benefit, a migration program by Azure focused on existing Windows users to migrate their infrastructure and Windows licenses to Azure. There’s a strategic utilization of Reserved VM Instances and sticking to Windows licenses that can yield significant savings over GCP or AWS.

Where cloud cost management comes in

Spinning up VMs on Azure: easy. Managing their costs at enterprise scale over time: ehh, it gets complicated. Fortunately, there are cloud cost management platforms that allow you to gain cost visibility, allocate to the right cost centers, and optimize with ease. Use a platform with a robust data and analytics engine that allows your teams to ingest cloud costs and utilization over time to:

• Inform engineers and finance teams of what they’re actually using on Azure and being billed for with easy-to-use reporting
• Increase optimization of these cloud costs by using a platform that helps create rightsizing recommendations so you use the correct size and types of Azure services that fit your actual business needs
• Improve operations by translating cloud costs and spend into strategic business choices—because cloud isn’t just a cost center—it’s a strategic asset

If you’re using Azure Compute services today and want to try a cloud cost management platform, you’re welcome to try Apptio Cloudability for free today.