How to choose the best virtual machine for your workload in Azure

Azure makes it much easier to choose a virtual machine type than other cloud computing providers. Azure’s VM packages offer a more simplified and compact experience than other providers. This makes it easier to make the right decision when you need to sell your business soul and technology to the sky.
However, it is still important to think carefully about choosing a VM instance for Azure. It is easy to overprovision services and, at the same time it is easy for resources to be underprovisioned. This can lead to strange problems during testing and application installations.
Today we’ll be going through the alphabet soup of Azure VMs. We’ll start by giving a brief overview of typical workloads, and which Azure VM instance to use. You can find more information about Azure’s VM types and examples of workloads in the Azure help section.
Which Azure VM should you use for Workloads?
Azure VMs come in a variety of configurations depending on the hardware and performance they support. We can help you choose the right Azure VM.
We know that sometimes you have limited time and need to find an answer. Here’s the tl:dr version of which Azure VM to use:
Type of VM
Use the Case
Tiny web servers (blogs with low traffic, low processing), testing, development, small to moderate databases, proof of concept, or MVP app staging
Tiny web servers (blogs with low traffic, low processing), testing, development, small to moderate databases, proof of concept, or MVP app staging
CRM systems, more extensive database, web-servers, web-servers, small production apps with medium web traffic, desktop virtualization and e-commerce, enterprise-level applications
SAP Hana, Netweaver large databases, data warehouse, data warehousing apps, analytical apps. Apps that perform heavy processing such as weather or financial apps. Apps that require heavy in-memory usage
Apps that require more processing speed include web servers, game servers and analytics servers.
Large databases, ERP solutions and SAP apps, data warehouse, data warehousing, and apps that use lots of memory are all examples.
For high-performance applications
Cited from Azure
Fluid dynamics, finite element analysis and seismic processing, reservoir simplification, risk analysis, electronic device automation, rendering, Spark weather modeling, quantum simulations, computational chemistry, heat-transfer simulation
High-performance databases and applications that require high storage throughput
SAP HANA, high-performance databases, massive parallel computing, large in-memory apps like Redis and Hekaton, High-performance databases
Large in-memory apps like Redis and Hekaton
Anything that uses a GPU (Desktop games VM, machine learning, etc
Okay. Okay.
Although choosing an Azure VM class may seem daunting at first glance, once you understand what each Azure VM tier offers, it becomes easy to choose a VM instance. So, let’s get started!
Let’s first separate Azure VMs according to their age. The most significant difference between Azure VM Tiers is the age of the hardware used to support them. The A, B and D series VMs use older hardware such as Broadwell and Ice Lake CPUs. The E, F and H series use newer Intel Scalable Xeon CPUs and AMD Epyc CPUs.
Each VM class has its own storage and memory types. The A, B and D Azure VM instances will have slower memory and slower mechanical or SSD storage drives. The E, F, H and L Azure VM instances will have faster memory as well as NVME drives.
Azure N class VM instances are a class apart. This Azure VM class includes GPUs. The Azure N VM instances run on Nvidia Tesla GPUs and do not have access the dedicated graphics processor. If you are looking to create your own cloud gaming services or need GPU acceleration, please visit the GPU Acceleration page.